Re: More on mutt losing its keymap
[EMAIL PROTECTED] writes: That's what I get for being a command-line freak: everyone else is using xterms and xterm termcaps seem to work fine between linux, irix and solaris. I guess I have to sit down and set up linux termcaps on our sun/sgi boxen -- using vt100 as I do now is clearly sub-optimal. Of course, another option would be to modify your linux terminfo entry on your machine so that terminfo treats your console similar to the way it treats vt100; clearly, the console will respond to those same codes and I'm a bit surprised that the console's terminfo entry is so different from vt100. The only reason I can see is that the ncurses maintainers decided that the Cursor/Application mode switching which vt* terminals are set up to do was Sick And Wrong (tm). However, changing the linux console's entries to the right thing while the rest of the world still used the mode switching leads to problems like this. Also, since you have made this change to your linux terminfo entries, there isn't really a problem with using vt100 control codes elsewhere. There might still be an issue with the numeric keypad; if you encounter it and it's a problem I'd append \e to the smkx and rmkx string definitions. (That'll force the keypad to be numbers or arrow keys, depending on numlock - I don't think that there are any applications that use the keypad as keypad keys and that go through terminfo entries). The more I think about it, the more I'm convinced that the linux terminfo entry should have smkx and rmkx strings that set the cursor keys properly.
Re: More on mutt losing its keymap
[EMAIL PROTECTED] writes: Hi all, I whined before about mutt losing Up and Dn keys seemingly at random. Well, it looks like it's not random after all: keymap gets screwed after I telnet/ssh to other boxen on the network. Just now it happened after I telnetted to an irix machine and that killed the session with ^] - quit. Any idea what's happening there? -- I'd like to get this fixed, it's very annoying when arrow keys stop working all of a sudden. Ok, here's my guess: Your cursor keys are getting stuck in Cursor Mode when they should be in Application Mode or vice versa. Briefly, vt100-like terminals (such as xterm or the linux console, or that old vt220 rotting in my parents' spare room) have two modes that cursor keys can be in - one in which the keys transmit escape[A, escape[B, escape[C and escape[D, and one in which they transmit O (that's a capital o) as the second character instead of [. (The reason for this has something to do with DEC marketing and the transition between the vt52 and vt100, I think) Anyway, terminals can be switched between the two modes simply by getting the appropriate escape codes from the other end. In an ideal world, there'd be only one set of codes transmitted by the cursor keys. In an almost ideal world, mutt would switch to Application mode on startup, switch back to Cursor mode if you spawned a shell or put it into the background, and then back again to Apllication mode once it either comes back to the foreground or the subprocess exits. (Or mutt would expect Cursor mode and switch there) Now, one of these things is not happening, and from your earlier posts my guess is that mutt expects the cursor keys to be in Cursor mode and just assumes that they'll be that way. I make this guess because each vc initially starts in Cursor mode and switches to Application mode only when told. ... investigations of mutt, slang and jed I won't bother you with [1]... Ok, here's what I think the end result is: 1. Cursor mode is the normal mode; if an application switches the keys to Application mode, it damn well better switch back. 2. jed can deal with both the Cursor and Application mode key sequences no matter what terminal it's on. I suspect that jed falls back on its own keybindings and isn't dependent on what slang tells it. 3. the Linux terminal responds to the vt100 codes that switch cursor state. These codes are NOT documented in the console_codes manpage. (though the related codes to switch the numeric keypad between numbers and escape sequences are documented) 4. For vt100 (and relatives), the terminfo entry says that the key sequence for up is Escape-O-A. The terminfo entry for the linux console says that the key sequence for up is Escape-[-A So I'm undecided as to who's to blame here. If all access to the console was done through the terminfo entries that Debian has, then all would be fine - no program would ever send the linux console the control codes to switch into Application mode, and every application would expect the console to be in Cursor mode. However, that's not what's happening here - something keeps switching your arrow keys into Application mode - probably this is some program on the far end that either has a terminfo entry for linux that says to do this or is using a vt100 definition for your terminal. I'll note that the terminfo entry for vt100 says that applications which expect to read in movement keys should, on startup, send the codes that switch the arrow keys to Application mode and cause the numeric keypad to send escape sequences. At the end of the application, that same terminfo entry says that they should send the codes that return the keypad to just numbers and the arrow keys to Cursor mode. As a temporary measure, if you've just disconnected and think that your cursor keys may be messed up, you can do TERM=vt100 tput rmkx Your keys should then work the next time you start up mutt. But what you probably want is a permanent fix. The most elegant solution I can come up with to fix this is to do the following as root: infocmp linux /tmp/linux.tic echo ' smkx=\E[?1l, rmkx=\E?1l,' /tmp/linux.tic tic /tmp/linux.tic rm /tmp/linux.tic This will ensure that any application which wishes to use cursor keys in linux at least has a chance of sending the codes to switch the console into Cursor mode in case something switched it out. (Yes, I do mean for both smkx and rmkx strings to be identical) You can also run the same sequence of commands as a non-root user, but that won't change the system-wide setting. If you want, before you get rid of /tmp/linux.tic, you can send it to that IRIX system and use tic there to teach it about the linux terminal. Perhaps this adjustment to the terminfo entries for linux could be sent upstream; I guess filing a bug against ncurses is the way to do that.[2] By the way, I found the page http://www.mpimf-heidelberg.mpg.de/~rohm/computing/mpimf/notes/terminal.html
Re: apt-get, upgrade, /var/cache/apt/archives
Jason Gunthorpe [EMAIL PROTECTED] writes: On Fri, 19 May 2000, Antonio Rodriguez wrote: 1:) What is the best way to make apt-get use the /archives folder to perform the upgrade and return the system to stage 2. above? cp /archive/*.deb /var/cache/apt/archives/ Specifically, apt stores all of the dependency and install-order information in databases that only need to download the Potato Packages.gz files to be rebuilt - if you do this, then do your update and dist-upgrade, apt shouldn't have to download packages again (except that some of the packages will have changed since you last got them). What this will not restore is any changes you made in which packages were selected. This information is stored elsewhere.
Re: /dev/fb not there?
Steve Hsieh [EMAIL PROTECTED] writes: You can use mknod to make them yourself. Ick. Inevitably, if you get used to doing things this way, you'll slip up at some point and create things with the wrong permissions. A much better solution is to use /dev/MAKEDEV to do it: /dev/MAKEDEV -v fb (this must be done as root). The point is that the existence of files in /dev has little to nothing to do with whether or not the kernel supports that device; creating the device files and supporting those devices in the kernel are separate tasks. (I do remember some talk of a kernel patch that would make /dev into a virtual filesystem like /proc, so that files would appear in /dev magically as soon as the kernel was made to support the given device, but that's not here yet)
Re: fetchmail, why does it do this?
Ingo Hohmann [EMAIL PROTECTED] writes: On Tue, Feb 09, 1999 at 12:01:16AM +0100, Helge Hafting wrote: Well, maybe fetchmail gives up completely at the first broken message? At least, thats what it does here. But debian.virtual.de.host is my own system, and I am able to _MAIL_ to [EMAIL PROTECTED], it appears in his spool file, only fetchmail won't get it back... One thing that may be relevant here is the fetchmail smtphost setting. By default, when fetchmail delivers mail it retrieves, the address it delivers to is the same as the local address that it was assigned when getting mail. An example: I have a box that connects via a ppp dialup line. When I connect, my machine is assigned an IP address in the range 128.220.222.*. For example, let's say I dial in and get 128.220.222.98. Then, when fetchmail connects to my mail server (128.220.2.5), it notices that the local end of the connection is at 128.220.222.98, and when it turns around and delivers the mail, it attempts to deliver it to 128.220.222.98. That is, fetchmail delivers the mail to my machine through the IP address associated with the ppp connection, instead of the localhost (127.0.0.1) address. (*) Now, since I don't trust the other people who use my ISP to not try to use my machine as a mail relay, my machine doesn't accept mail from IP addresses in the 128.220.222.* range. (In fact, my machine doesn't accept mail from anything but my own mini-network) Therefore, under the defaults, fetchmail will fail to be able to deliver mail since the fetchmail connection will appear to be coming from the IP address that my machine got when it dialed in. The default debian exim setup (which is, in my opinion, a good way to do things) refuses to allow non-local machines to deliver mail by just sending to the username, not to [EMAIL PROTECTED] The fix for this is to include the line: defaults smtphost localhost at the top of your .fetchmailrc file. This forces fetchmail to deliver mail to 127.0.0.1, which means that it will use the loopback network interface, which is guaranteed to be local. This means that my machine accepts the mail, and also means that under debian's default exim setup, delivery will be allowed to unqualified names. (*) delivery to localhost used to be the default. Why fetchmail's upstream authors changed the default (with only a small note hidden deep in the documentation) is completely beyond me - I can think of only very unusual reasons why you'd want the current default behavior, and many reasons why you'd want the old default behavior. Fetchmail's upstream authors have a history of introducing optional features which, while useful in some cases, will cause many installations to break, and then of enabling those features by default. I believe this to be evidence that the fetchmail author has taken his own writings on the benefits of bazaar-style development too seriously to impose enough sanity-checking control on fetchmail. (For those of you who don't know, Eric S. Raymond, fetchmail's upstream author, is the author of The Cathedral and the Bazaar, which many consider to be the manifesto of the open source movement. Fetchmail is in fact the program used as a case study in that article) For further rantings about this and other tendencies in the development of fetchmail, see debian-devel for the past few days.
Re: Removeing N lines from a file
Lance Hoffmeyer [EMAIL PROTECTED] writes: How can I use gawk or some other program to remove a number of lines from a text file. Initially, I only need to delete the top 10 lines from a file but it might be useful to know how to delete lines from any part of the file. The top 10 lines from each of these files vary in what they may contain so I need to indiscriminately delete them. I figured gawk is what I would need to use to perform this task but if someone ones of something else that might do this that would be fine too. Well, someone already mentioned sed, but why not just do: tail +11 infile outfile (The + tells tail to start printing things at line 11; to get the last 11 lines you'd use tail -11) For similarly removing the end of a file, see head.
Re: A downright wierd Netscape problem
Curt Daugaard [EMAIL PROTECTED] writes: When trying to open certain documents in Netscape--for example, the Printing-HOWTO at sunsite.unc--the document is not displayed right away but starts to download. When it's done a dialog box appears with this message: Netscape: subprocess diagnostics(stdout/stderr) Vim: Warning: Output is not to a terminal [24; 1H [0;10mVim: Error reading input, exiting... Vim: Finished. [24;1H Vim: Warning:Input is not from a terminal I have no idea what brought this on. Does anyone else? I'd be grateful to anyone who can give me a clue. Well, I don't know what brought it on, but I have an idea of how to fix it - go to the Edit-Preferences menu in netscape, then go into for Navigator-Applications (you may have to click on the little triangle next to Navigator). Then, select Plain Text in dialog box (it's likely all the way at the bottom). Notice that it's not set up to be handled by Netscape. Edit the type and select the Navigator option button in the Handled by... group. My; describing how to do something in a GUI is painful - it's so much easier to say change line 432 to... in the such-and-such file.
Re: docbook-stylesheets
[EMAIL PROTECTED] writes: Folks, I'm trying to use the docbook-stylesheets (v. 1.07-1). I've got a simple outline and I've managed to get jade to accept it (after a little struggling). I have two big problems though: When I try to use HTML output, I don't get anything generated (or if I do I can't find it). Can anyone tell me why this might be? When I use TeX output, I get a fine looking .tex file, but there's no instructions on further processing it: tex foo.tex doesn't work :) Can anyone help me with that? I don't know about the html output, but the tex needs to be processed not with regular tex but with jadetex (which is tex with a load of extra macros added to it) Install the jadetex package.
Re: minimum X packages?
Paul Nathan Puri [EMAIL PROTECTED]@office.law-counsellor.com writes: I recently upgraded to potato. Now X starts and immediately dies. Why would this be? What are the minimum packages necessary to run X under potato (or in general)? It would probably be because the X maintainer left a typo in /etc/X11/Xsession that means X starts and immediately dies. About lines 45-51 of /etx/X11/Xsession, change: fi fi default) ;; *) program=$(which $1) to: fi fi ;; default) ;; *) program=$(which $1) (notice the extra line with ;; in it). Methinks the X maintainer was releasing packages after having not slept sufficiently the night before (at least, when I have errors like this in fvwm95 packages, that's what the cause is).
Re: apt
Robert Kerr [EMAIL PROTECTED] writes: Is there a way to make apt not delete the .deb files after it installs them? -bob I assume you mean when using apt from dselect; otherwise, apt doesn't delete them automatically. There's no easy way. The non-easy way is to modify /usr/lib/dpkg/methods/apt/install and change the line which says: apt-get clean echo Press enter to continue. read RES exit 0; to: (true || apt-get clean ) echo Press enter to continue. \ read RES exit 0; (Oh I suppose you could just remove the reference to apt-get clean entirely and get something like echo Press enter to continue. read RES exit 0; but then how would you remember what you had done and how to undo it?)
Re: GNUcash/Libs
Will Lowe [EMAIL PROTECTED] writes: libXm.so.1 = not found This is a motif library. Some versions of Lesstif will work. libXmHTML.so.1.1 = not found XmHTML doesn't exist as a debian package. Read the GNUCash readme and get the source. I've been working on packaging Gnucash (which means packaging XmHTML and nana, also), but I can't get the gnome version to build, which is really the only one I'm interested in. First off, nana is already packaged. Secondly, although xmhtml isn't packaged, gtkxmhtml is packaged, and that's used for the gnome build. Thirdly, and most importantly, the gnome ui code is a big mess - even if it were to somehow miraculously compile (I don't know how those involved were ever able to build the -gnome RPMS), most features wouldn't work. The gnome code really isn't ready for anything but having some coders attack it - if anything compiles, that's just a testament to the fact that compilers can put up with a lot. That said, it seems to work well with lesstif 0.86.9 (the version in slink), although for that you would have to package XmHTML first.
Re: diald is eating packets
Kenneth F. Ryder III [EMAIL PROTECTED] writes: SNIP 1) how to keep diald from eating the first packet it sees coming across the dummy link, and instead hold it and send it down the PPP link once established. I think diald does this with UDP packets - with TCP packets there are other rather annoying issues/problems. However, the kernel's ipmasq routines do it, so you'd think there'd be a way. 2) how to increase the time limit in the standard filter (I'd like to use standard.filter over my very primitive one) on the FIRST packet sent out by a program (like telnet, rlogin, etc.) Well, not quite that, but what I added after a somewhat similar problem to my /etc/diald/diald.options is: impulse 30,0,0 This causes the link to always be kept up for at least 30 seconds.
Re: GNUcash/Libs
Timothy Hospedales [EMAIL PROTECTED] writes: Hello! I am trying to get GNUCash to work; the binary distribution (1.1.23); converted with alien from rpm gives libXm.so.1 = not found libXmHTML.so.1.1 = not found libreadline.so.3 = not found Can anyone tell me what packages I need to get these libs? Well, I don't know about that libreadline.so.3 (that looks weird - I didn't think there was a readline with that version number), but those other two libraries look like motif libraries - you can get libXm.so.1 from lesstif, but I don't know where to get the other ones. Why don't you try the gnucash-gnome rpm? That should be liked with only libs from the gnome project, and you should be able to either install it immediately or find the libs you need easily in Debian packages.
Re: doom (sorry)
Frankie [EMAIL PROTECTED] writes: Daniel Martin wrote: Frankie [EMAIL PROTECTED] writes: I have just d/l doom shareware version, unfortunately, it wants libXt.so.3 for the X version, and the svga version wants libc.so.4. What packages are these in? I had a look in oldlibs, but I'm not entirely sure what I should look for, so O didn't find anything. You want the xlib-compat package. I believe that this package has finally disappeared from Debian 2.1, though, so you'll need a 2.0 (hamm) archive or CD. I have the cheapbytes 2.0 CD, and that doesn't seem to have an xlib-compat package on it. :-( Any ideas where else I might look? Sorry; It's called xcompat, but you're right - I can't find it in 2.0 at all. You may have to go back to 1.3.1 to find it. Assuming you can't find someone with a 1.3.1 CD, you can get the xcompat and libc4 packages (which I just pulled off my bo CD) from: http://jhunix.hcf.jhu.edu/~dtm12/ md5sums: c3c4652f43e110de58fe4cec2a22771d xcompat_3.1.2-4.deb e161434c2d952ffce8c286d9343ff45a libc4_4.6.27-15.deb Unfortunately, I can't include the sources to these packages, as I happen to have... misplaced my bo source CD. Why, you may ask, were these packages dropped from Debian? Well, basically, because a.out is such an ancient binary format that it should really be forgotten, and all of the support tools necessary to make it work make continuing support of a.out format a somewhat onerous burden. Besides, these packages can't be built with the current Debian tools, so it would take some amount of work to bring them up to the state where they could be included again - it's not merely a matter of taking on a whole load of rarely-used packages. That said, I've been wondering why one of those emulators we've got floating around couldn't be made to run libc4 a.out format... pgptakQNkhqBw.pgp Description: PGP signature
Re: doom (sorry)
Frankie [EMAIL PROTECTED] writes: I have just d/l doom shareware version, unfortunately, it wants libXt.so.3 for the X version, and the svga version wants libc.so.4. What packages are these in? I had a look in oldlibs, but I'm not entirely sure what I should look for, so O didn't find anything. You want the xlib-compat package. I believe that this package has finally disappeared from Debian 2.1, though, so you'll need a 2.0 (hamm) archive or CD.
Re: Soundblaster under Debian 2.0
Ramesh Natarajan [EMAIL PROTECTED] writes: Hi, I just installed Debian 2.0 and compiled kernal (2.0.34) with Sound as module. I seem to get something working, but not all. I configured my /etc/isapnp.conf, did a isapnptools /etc/isapnp.conf followed by $insmod sound trace_init=1. On dmesg, I see Sound initialization started Sound Blaster 16 (4.16) at 0x220 irq 5 dma 0,5 Yamaha OPL3 FM at 0x388 Sound initialization complete when I do cat /dev/sndstat, I see Card config: Sound Blaster at 0x220 irq 5 drq 0,5 (SB MPU-401 irq 1 drq 0) OPL-2/OPL-3 FM at 0x388 drq 0 Iam able to play CD using workman. When I cat some .au to /dev/audio, (or when I try saytime) I get no sound. I get no error either. When I do rvplayer welcome.rm, I get some error like Codec not installed/Compression not supported. From what I see in HOWTO, looks like I have MPU stuff not enabled. Is this a problem? I donot have MPU 401 support enabled in make config (going by help text) I used to have RedHat 5.1 with sound working like sweat (using sndconfig). Is there a similar tool under debian? (just wanted to be a GNU purist for a while :) No; unfortunately there isn't such an easy sound configuration tool (anyone want to take apart RedHat's? - and while we're at it, compile sound as a module in the default kernel?). However, I have sound working on my machine, on my PnP Soundblaster clone. I too get the MPU thing enclosed in ()'s, so that in and of itself isn't your problem. What does the rest of cat /dev/sndstat say? (specifically, does it list an audio device? What does it say about installed drivers?) Add a NAME section to your isapnp.conf right before the (ACT Y) thing, like this: (NAME SoundCard) (ACT Y) ) And then make certain that the VERBOSITY setting at the top of isapnp.conf is at least 2. This should make isapnp spit out information about the card after configuring it - make certain this information matches what your kernel thinks about your card. Thats my main Qn. I have few more :) 1) I can do startx only as root. Not as any other user. I get some error like xlib: conncecion refused by server. If I run xdm, I can log on as any user though. This one is easy. Look at the file /etc/X11/Xserver. Read what it says. 2) When I try make xconfig for kernal config, I cannot select any of the y/m/n options. I can input all text fields (like irq, IO) though. Wierd.
Re: [offtopic] Dont't ECHO
Nuno Carvalho [EMAIL PROTECTED] writes: Hi, I'm trying to, using socket programming, ask the remote user for login and password. I would like that password won't appears on remote machine when user is typing it. I already saw RFC's but it doesn't worked ! :((( Could someone send me such part of code !? Should IAC + DONT + TELOPT_ECHO sent to socket or client descriptor !? You're misunderstanding the RFCs. In terms of telnet options, the ECHO refers to whether the local or remote end of the connection should handle echoing the characters back to the user. The way telnet does what it does is by having the server send IAC WILL ECHO and then having the server only echo those characters that should be echoed. I have attached below a transcript of a simple telnet session to show you what's going on. You may also need to send the telnet option that means one character at a time transmission and not line-by-line mode. I'm not quite certain how to do this; I believe it has something to do with manipulating the go-ahead option, and supressing it. I'd try to do the following: Server IAC WILL SUPRESS-GO-AHEAD Client IAC DO SUPRESS-GO-AHEAD Server present login prompt Client send username Server IAC WILL ECHO Client IAC DO ECHO Server present password prompt Client send password Server IAC WONT ECHO I'm not certain what you should get if the client isn't willing to do the supress-go-ahead stuff; if the client responds with a don't echo, then there's really no way to prevent the password from being displayed. Looking at your mail again, it appears that you are writing both the client and the server part of this application - if so, then you shouldn't be touching telnet options at all - you should only deal with the telnet options if you expect people to be using telnet to connect to your server. If you are writing the client as well as the server, look at the getpass function that someone else mentioned, or look at the termios(3) manpage for how to turn terminal echoing off. Here's the transcript of a telnet session to my machine - the CLIENT or SERVER refers to the machine which sent each bunch of data. Note that this transcript is complicated by many other options, but I think that only the echo and supress-go-ahead options are relevant. (These begin around record number 5, below). Record number 1 (SERVER), length 12 FFFD18FF FD20FFFD 23FFFD27 .ý..ý .ý#.ý' 01 Record number 2 (CLIENT), length 12 FFFB18FF FB20FFFC 23FFFB27 .û..û .ü#.û' 01 Record number 3 (SERVER), length 18 FFFA2001 FFF0FFFA 2701FFF0 FFFA1801 .ú ..ð.ú'..ð.ú.. 01 FFF0 .ð 02 Record number 4 (CLIENT), length 38 FFFA2000 33383430 302C3338 343030FF .ú .38400,38400. 01 F0FFFA27 00FFF0FF FA180068 79706572 ð.ú'..ð.ú..hyper 02 7465726D FFF0 term.ð 03 Record number 5 (SERVER), length 15 FFFB03FF FD01FFFD 1B05 FFFD21xx .û..ý..ý..û..ý! 01 Record number 6 (CLIENT), length 24 FFFD03FF FC01FFFB 1A1F 00500018 .ý..ü..û..ú..P.. 01 FFF0FFFD 05FFFB21 .ð.ý..û! 02 Record number 7 (SERVER), length 43 FFFB0144 65626961 6E20474E 552F4C69 .û.Debian GNU/Li 01 6E757820 736C696E 6B206375 73682E64 nux slink cush.d 02 796E2E6D 6C2E6F72 670D0Axx yn.ml.org.. 03 Record number 8 (CLIENT), length 3 FFFD01xx .ý. 01 Record number 9 (SERVER), length 14 0D0A6375 7368206C 6F67696E 3A20 ..cush login:01 Record number 10 (CLIENT), length 1 75xx u01 Record number 11 (SERVER), length 1 75xx u01 Record number 12 (CLIENT), length 1 73xx s01 Record number 13 (SERVER), length 1 73xx s01 Record number 14 (CLIENT), length 1 65xx e01 Record number 15 (SERVER), length 1 65xx e01 Record number 16 (CLIENT), length 1 72xx r01 Record number 17 (SERVER), length 1 72xx r01 Record number 18 (CLIENT), length 1 6Exx n01 Record number 19 (SERVER), length 1 6Exx n01 Record number 20 (CLIENT), length 1 61xx a01 Record number 21 (SERVER), length 1 61xx a01 Record number 22 (CLIENT), length 1 6Dxx m01 Record number 23 (SERVER), length 1 6Dxx m01 Record
Re: refused connect from 'unknown'
Pere Camps [EMAIL PROTECTED] writes: Hi! Can somebody explain me what this is? Dec 7 13:52:11 casal in.telnetd[27798]: warning: can't get client address: No route to host Dec 7 13:52:12 casal in.telnetd[27798]: refused connect from unknown If my machine has a telnet request, then my machine knows the IP (at least) of the machine which requests it, no? No - not if the person connecting disconnects almost instantly; what can happen is that if the person in question opens and then closes a connection almost instantly, the connection goes to inetd, which accepts it, but before tcpd (which is what inetd hands telnet connections off to, and which is the program generating these log messages) gets the connection and finds out who's on the other end, the connection is closed, and tcpd is left without a clue, hence the confusing error messages. This is usually done as part of a port scan - testing to see which ports are accessible on your machine. There ought to be an option to inetd to log all tcp connections before passing them off to something else to handle, but I can see how that could get to be a hassle on a busy machine. On the other hand, services which are not run from inetd - for example, apache on most machines - will know where this connection was coming from, and many port scans hit port 80 as well as port 23. I seem to remember some program that monitored every individual incoming network packet and logged warning messages about suspicious packets - I suppose someone will know how to do this with ipchains or ip firewalling stuff.
Re: No more C++ ?
Michael Beattie [EMAIL PROTECTED] writes: On Tue, 1 Dec 1998, Ossama Othman wrote: Hi, robinson:~/uni/c++/src$ c++ hello.C /usr/bin/ld: cannot open -lstdc++: No such file or directory collect2: ld returned 1 exit status I heard some people have the same problem. But how can I fix it? You need libstdc++, install it. That's not the solution, at least for me. I _had_ working C++ prior to the __register_frame_info problem upgrades. I checked /usr/lib. All of the usual C++ libraries are there, including libstdc++-2.8 and 2.9. There is a thread in -devel about this, a problem with a link not extisting or something, check the recent archive for something on it. Really? I couldn't find that thread. In any case, the problem is that there's no file libstdc++.so - I fixed this by (as root) doing the following: cd /usr/lib ln -s libstdc++-2-libc6.0-1-2.9.0.so libstdc++.so There will be a new libstdc++2.9-dev soon enough.
Re: ANSI Color Escapes in $PS1.. heh.
[EMAIL PROTECTED] writes: How would one check to make sure the terminal is capable of ANSI escape sequences? -brad I'd do this by not coding in the escape sequences directly, but by using tput: export PS1=\[`tput setaf [EMAIL PROTECTED] sgr0`\]:\[`tput setaf 4`\]\w\[`tput sgr0`\]\$ tput is a command-line interface into the terminfo database - if your terminal is known to be able to handle colors, it'll generate escape codes and if not it'll generate nothing. Thus, this prompt line works just as well on a color-capable xterm as it does on an old Wyse terminal (which not only can't do color, but barfs if handed VT-style escape codes).
Re: undefined symbol: __register_frame_info
Eric Jensen [EMAIL PROTECTED] writes: I recently installed Slink on a laptop. I have the latest versions of all the signifigant packages. When I try to run dselect, groff, or many many other things, I get this: error in loading shared libraries /usr/lib/libstdc++-libc6.0-1.so.2: undefined symbol: __register_frame_info How do I fix this??? You complain to upstream maintainers to fix their packages. Until they do, you can use an older version of libc6. I've put the last one I had at: http://master.debian.org/~fizbin/libc6_2.0.7u-3.deb Why did this happen? Well, there were a few versions of libc6 which got out which were broken. The way they were broken meant that programs compiled with them (not C, just C++, pascal, and anything else using certain egcs compilers, as opposed to gcc stuff) couldn't be run anywhere else - they all had this weird dependency on this odd symbol. Now, once libc6 is fixed, we have this scenario. In any case, downgrade libc6 until all of your other packages have been upgraded and this stupid libc6 inconsistency is fixed.
Re: Linux shutdown
Tun Yang [EMAIL PROTECTED] writes: Hello ... I was wondering if there was a way for linux to make use of the ATX soft power switch... For example, shutdown ... totally, power and all just like win9x. or, suspend to disk by pressing the momentary button? Recompile your own kernel (install the kernel-pkg package and either download a linux .tar.gz source from ftp.kernel.org or install one of the kernel-source packages) and, when choosing options, enable APM (which stands for Advanced Power Management).
Re: diald time restriction
Peter Bartosch [EMAIL PROTECTED] writes: hi debians i´ve got only a short question: is it possible to restrict diald only to work at e.g. one hour per day? Yes. and how? (config file, cron-job, etc.) Whatever. You can restrict certain portions of the config file to be operative only at certain times of the day, you can have a cron job completely shut diald down and start it up at specific times. It depends what you mean by work - do you mean only bring the link up if receiving network activity during certain times of the day, (and keep the link down unless specifically requested at other times) or do you mean completely give up managing the link (both in terms of bringing it up and taking it down on inactivity) during certain times of the day? Either situation could be done with a config file change, though in the second situation it's better to use a cron job. i don´t use diald at the moment! it´s only to avoid reading more and more ,-)
Re: DEL key in Xemacs
Daniel Elenius [EMAIL PROTECTED] writes: I have this problem that has been bugging me for quite some time: I can't get my DEL key working properly in Xemacs. I want it to always generate delete-char, but it works like backspace, i.e. generates delete-backward-char. I have a swedish keyboard. I've put (require 'iso-syntax) in my .emacs. One of the most wonderful things about xemacs20 is the extent to which it is easily customizeable through the custom.el interface. Do M-x customize - when asked for the group say editing-basics. One of the choices will be Delete key deletes forward - it is off by default; change it, then make certain you both Set and Save the settings in that group. Then, it should work as expected (maybe after a restart of xemacs).
Re: Ifconfig
Amanda Shuler [EMAIL PROTECTED] writes: I have a question about ifconfig. I have a machine that I am trying to configure to put onto a local network. I'm an assigning it IP address 192.168.76.76 I type: ifconfig eth0 192.168.76.76 then I check it with ifconfig and everything is correct. I reboot the machine, and recheck ifconfig -- it's wrong. It resets the IP address to 192.168.1.1 everytime! Currently, I do not have this machine physically hooked up to the network, because I was just doing the configuration and I didn't want to knock another (very important) machine off the network. If the ethernet card is not actually hooked up to the network, will that cause this reset to happen upon every boot? How do I get it to stay at 192.168.76.76? ifconfig affects something only so long as the machine isn't rebooted - the ethernet card itself never knows what IP address it has; only the kernel knows this. Therefore, at system startup, the kernel is told by the intitialization scripts what IP address the ethernet card has. This is done in the script /etc/init.d/network - go edit that file to suit your new IP address. (That is, change any ifconfig line that's setting the 192.168.1.1 ip address, or, if there is no such line, add the ifconfig line you use above to that file).
Re: How to enable DPMS monitor poweroff?
Rick Macdonald [EMAIL PROTECTED] writes: The Hardware-HOWTO just says: 20.1. VESA Power Savings Protocol (DPMS) monitors Support for power savings is included in the Linux kernel. Just use setterm to enable support. but setterm insists: timshel:~$ setterm -powersave on cannot (un)set powersave mode I can't find any more details anywhere. I have an ATI mach64 (Grpahics Pro Turbo) and a Viewsonic 17GS monitor. Kernel 2.0.32. setterm is for modifying stuff on the linux console, not from X. You're probably trying this from an xterm. Inside X, use the 'xset' command, like this: xset dpms -or- xset dpms 300 600 900 You might consider putting this command into /etc/X11/Xsession, after the bit about xmodmap. If you use xdm, you might consider putting this command into /etc/X11/xdm/Xsetup_0 as well. I also find 'xset dpms force standby' a nice command when I want to go to sleep. More info on the xset manpage.
Re: more X in slink stuff...
Zack Brown [EMAIL PROTECTED] writes: I'm using slink. When I run startx, I get the following errors: xauth: error in loading shared libraries libXmu.so.6: cannot open shared object file: No such file or directory xauth: error in loading shared libraries libXmu.so.6: cannot open shared object file: No such file or directory xinit: error in loading shared libraries libXmu.so.6: cannot open shared object file: No such file or directory Aside from the fact that X is broken in slink, is there any way around this? What deb file is the libXmu.so.6 library available? I think this may be another case of the two xlib6g 3.3.2.3a-7 packages causing problems. The basic story (as I understand it) is that there was a broken xlib6g 3.3.2.3a-7 that got released accidentally - although it was very soon replaced with a working one, some people got bitten by the bad one. At least, I think that's what happened. I've also heard things to the effect that the problem is caused by a bad pre- or post-rm script in the -6 xlib6g. In any case, go get xlib6g_3.3.2.3a-7.deb again from your local mirror, make certain that it has md5sum 276d3782f9be91508241592902146e49. Reinstall it and this error should clear up. pgpDR278MBojZ.pgp Description: PGP signature
Re: dual boot Linux/NT question
Damir J. Naden [EMAIL PROTECTED] writes: Hi Rakesh Mohan; unless Mutt is confused, you wrote: SNIP BTW, I can now use lilo for branching to NT and linux, this works fine. That is interesting. I was never able to do this. Maybe because my NT partition was NTFS.. damir I just today set my system up to do this - and my WinNT partition is indeed NTFS. What I did was make a linux boot floppy, install NT (onto a system that had had Debian on it for ages), boot linux, and then did: dd if=/dev/hda of=/boot/ntmbr bs=512 count=1 This saved my NT MBR into a file, which I then reference in my lilo.conf, and then I re-ran lilo. For reference, here's my lilo.conf: boot=/dev/hda root=/dev/hda1 compact install=/boot/boot.b map=/boot/map vga=normal delay=20 image=/vmlinuz label=Linux read-only image=/vmlinuz.old label=old read-only other=/boot/ntmbr label=WinNT table=/dev/hda This isn't ideal (lilo doesn't really start NT, but NT's boot menu), but it works. Now, as I've forgotten my NT password, I'm going to have to reinstall it (unless someone here knows the magical files to erase/modify so I can change my NT administrator password from Linux) so I'm going to get to go through this again...
Re: /etc/host.deny and co
Linh Dang [EMAIL PROTECTED] writes: Thank you very much! Another question if you don't mind? Someone mentionned ipfwadm. What do you think about it, how does it compare to tcpwrapper? Does one has to recompile the kernel to use ipfwadm ? Yes, your kernel has to be compiled to support ipfwadm; no, I don't know whether or not the default Debian kernel is. The basic difference is where they operate. tcpwrappers operates at the socket level; it gets invoked after a network connection has been made but before that connection is passed off to the actual program that does telnet logins, or accepts mail, or ... ipfwadm puts its blocks in at the kernel level, so that packets trying to establish network connections you don't want never make it through to the kernel logic that would establish a connection. ipfwadm is most useful when your box is acting as a router, and you wish to protect machines on one side of the network from machines on the other side. But it can also be useful in your case. As for which is more secure - ipfwadm is certainly the one to use for the ultra-paranoid. It is possible that a SYN-flood type DOS attack (an attack where some malicious person tries to initiate as many connections as possible in rapid succession - the idea isn't to break in, but just to bog down your machine and so make your life miserable) could get through on a tcpwrapper-protected machine and be blocked on an ipfwadm-secured machine. However, since you are leaving port 80 (http connections) open anyway, the attacker would just have to target that port in their SYN flood. Also, in my case my machine is just connected through a phone line, and so packets of any kind can only reach my machine comparatively slowly. tcpwrappers provides for more extensive logging of what's going on in my experience; I have this silly idea that some day I'm going to get to file a CERT report because some hacker who'd hacked their way across many systems wound up in my logs. Hasn't happened yet, but you never know... By the way, that hosts.deny line I use is now: ALL: [EMAIL PROTECTED] : rfc931 : spawn ((echo %c %a contacting %d; /bin/netstat --inet -n; traceroute -p 31434 %a) 21 | mail root) The echo and dumping to a file in /tmp were earlier debugging features I meant to change but had never gotten around to.
Re: conflicts in Debian Distributions
[EMAIL PROTECTED] (Lance Arsenault) writes: I just would like to know if Debian 2.0 has conflicting software in it like Debian 1.3.1 . Background: Debian 1.3.1 would not let you install all the software in the release. Installing all the software in the release saves a lot of time, and hard drive space is cheeper than time. Yes, Debian 2.0 also has this feature. No, it is NOT a problem. While it may cause (a very small amount of) extra thought at install time, this is really a benefit to the user. Besides, the 2.0 setup is streamlined so that common configurations can be selected easily at install time. I have used Debian 1.3.1 in the past but I think that having packages that confict is a bad thing. I think that making all the packages in Debian compatable would be a big plus. In Debian 1.3.1 this is not the cast. For example you cannot install emacs and xemacs in Debian 1.3.1 . I'm guessing some of the filenames in these two packages are the same. To get rid of this confict you can just install them in different directories or something like that. Well, what do you mean by making all the packages in Debian compatible? We would end up with a totally unuseable system. What do you propose we do with, say, the multiple mail delivery programs? Should I be able to install sendmail, exim, smail, qmail, etc. and have my machine magically know which one is in charge of getting mail? Better to have the packages conflict, so that I can make the choice through the install program, rather than hunting down and disabling the programs I don't need. It's a pain to have to pick through 1000 + packages to install. I prefer to just install all of them without picking through them. Hard disk space is now cheep, and time is not. But by installing everything at once you just move the time burden around; you don't eliminate it. You still will have to make decisions about which package you use for which task. Now, perhaps your complaint about install time has to do with the incredible slowness of the dpkg-cd access method. This I can sympathize with - hunting through all the packages to find new ones is potentially more reliable, but much too slow. Try adding your CD drive to /etc/fstab and using the dpkg-mountable method - it is much faster. The basic problem of conflicting packages is a direct consequence of the choice Debian offers to users by providing all these alternatives. I would rather have a choice than not. What Debian should do is have a standard list of packages which comes preselected right after the install, and which the user could tweak before the first install run in case they wanted to add something initially. This standard set shouldn't have any internal conflicts, so that people can just install the standard set right away. But wait! We already _do_ do this. What might be nice is having some help during installation that forced users to realize that install everything is not a sensible option. This is not spelled out as loudly as it should be.
Re: root - password wiped ?? how ??
Richard E. Hawkins Esq. [EMAIL PROTECTED] writes: Thanx but nope, thats not the problem, sysklog 1.3-30 is installed and permissions Any other takers ;0) I'll take another swing :) Could something have switched you to/from shadow passwrds? This is another popular way to loose your root password :) If you're not using shadow passwords, just boot from a rescue disk, edit the password file, delete the root password, and reboot. You will have no password for root, and can now login as root and rest it. I assume it's possible to eidt your password frile from a rescue disk if you have shadow passwords, but it bats me how. Maybe delete the passwd for root in /etc/shadow? You just need to delete the x in the password field in /etc/passwd - that causes /etc/shadow to not be used for that account.
Re: Pine Attachment Associations
Stephen A. Witt [EMAIL PROTECTED] writes: I'm using Pine 4.05 on a Debian 2.0 platform at work. MS Office is the standard desktop automation application around here except for a small band of Unix enthusiasts. I've just installed Star Office, mostly so I can read the MS Word documents everybody sends me. I'm trying to reconfigure Pine so that it will automatically start up swriter3 for a attachment type of DOC. I thought the way to do that was to change /etc/mime.types, which I did. I modified the line that specified msword as the application for doc and dot files to be swriter3, but Pine still thinks its an MSWORD application. I also looked in the /etc/mailcap file but didn't find any mention of msword in there. I've looked on the Pine Info site and it mentions this and states that one should change the /etc/mime.types file. Anybody have any experience with this? Thanks... Changing the /etc/mime.types changes what type will be associated to a given file, but it doesn't say how to handle it. That's what /etc/mailcap is for. I can only assume that the pine info site assumes one already has pine configured to handle swriter3-type documents. In any case, changing /etc/mime.types won't help much for those attachments which are sent to you already labeled with a type. (really, most attachments should come with a content-type header). What you should to do is change /etc/mailcap to tell pine how to open ms-word files, and change /etc/mime.types _back_ to saying application/msword for doc files. To change /etc/mailcap, add the following (all as one long line) to the end: application/msword; swriter3 %s; test=test $DISPLAY != ; description=M$ Word document; nametemplate=%s.doc Since it appears that you're getting documents labeled with the type application/MSWORD you may also want to add: application/MSWORD; swriter3 %s; test=test $DISPLAY != ; description=M$ Word document; nametemplate=%s.doc to /etc/mailcap as well; I don't know if mime types are case sensitive.
Re: Removing ^M in files--in bulk?
[EMAIL PROTECTED] writes: I have a bunch of files (est. 200) which were brought over from OS/2 after being detached from emails (I've not got Debian networked yet here at work--subject of another post). All of them have this control character (^M) at the end of each line, as seen in vi (which I know v. little about except very basic I/O). These need to be removed before the files can be compiled. I am really hoping there is a way to do this in bulk ... using sed or something similar?? I am v. new to Linux, so have no grip yet on the more powerful utilities and/or syntax. (Nor am I a programmer.) Is there a way to do this?? Thanks! Install the sysutils package and then you can use the dos2unix command (also known as fromdos) - it does bulk ^M stripping.
Re: legality of mailing Debian CDs from US to overseas?
John B. Fink [EMAIL PROTECTED] writes: SNIP Now -- my question is, can I get into any trouble by exporting CD-R of Debian Linux? I'm just sending burns of the binary-i386.raw images that I get via rsync. My primary worry (all of a sudden -- I guess I should've thought of this beforehand but I was *pretty* sure it was okay) is strong crypto and the (imho, half-asswd) U.S. munitions export laws concerning crypto over 40bit. This is why there exists a non-us section which is distinct from the regular main, contrib, and non-free sections. By design, binary-i386.raw does not contain any export-restricted cryptography. I just want to make sure I can continue my Debian Giveaway project without fear of the FBI or some mysterious foreign government equivalent pounding down my door. You're sending CDs to the PRC? They have some odd censorship laws there - all that talk about freedom in free software might be risky :-). (Not to mention some places with strict censorship laws might not like the fortunes-offensive package) But seriously, as long as the FBI behaves intelligently, they won't be knocking on your door (for this at least). That said, I'm uncertain what the US laws are with regard to non-commercial (you're giving the CDs away) export to countries (such as Cuba, Iraq, Libya, Sudan, North Korea, Iran, and Syria) against which the US currently has an embargo. If you are really paranoid, you may wish to avoid sending CDs to those countries. Then again, the Iranian grad. students here use Netscape all the time and no FBI people are swarming over the campus computers, so...
Re: can any user have the right of root and can lunch Netscape
Jim Foltz [EMAIL PROTECTED] writes: On Sat, Oct 03, 1998 at 09:31:53AM +0800, Chan Min Wai wrote: Hai, I'm asking about that can a user have the rithr of root and can lunch netscape as well in Xwin.. Any user can start X windows if it is intalled properly. Any user can start netscape if it is intalled properly. Root user is not needed to run X windows or netscape. Do you have X windows and netscape installed properly? I think the question is whether one can run netscape as root. Currently, the debian netscape script prevents this (for good reasons, I think). It's possible to run netscape directly (bypassing debian scripts) by doing something like: MOZILLA_HOME=/usr/lib/netscape /usr/lib/netscape/netscape However, I'd _strongly_ recommend against it. Much better (if you need to run netscape, but you're logged in as root) is to do: su -c xauth add `xauth list $DISPLAY` dcmwai su -c netscape dcmwai Instead of dcmwai you should use your non-root username. This runs netscape as dcmwai, not as root, but displays it on your screen. Another possibility (which is what I do) is to never log in as root, but use su or sudo whenever one needs to do something as root.
Re: access to iso9660 raw image
Eugene Sevinian [EMAIL PROTECTED] writes: Hi All, Today a friend of mine gave me a CD with raw binary image of hamm on it. MD5 sum is correct. Would it be possible to use this CD as a hamm distribution without burning new CD. May be using loopback device or something like that? Any help is greatly appreciated. Yes, though of course it won't be possible to boot from the CD. Just do: (assuming the cd is already mounted under /cdrom, and the file is main.raw) mount -t iso9660 -o loop,ro /cdrom/main.raw /mnt Then, assuming this works, the best thing to do is probably to create a specific directory to mount it on (say /mnt/debiancd - of course, you need to umount the cd image first) and then place a line like this is your /etc/fstab: /cdrom/main.raw /mnt/debiancd iso9660 loop,ro,noauto 0 0 Now you can mount it with just mount /mnt/debiancd. Then, use dselect's mountable access method - I suggest this because it's faster than the mounted method and accessing things through the loop device is going to slow things down a bit to begin with.
Re: Free debugger that can do source debugging without executable.
Christopher Barry [EMAIL PROTECTED] writes: Is there a debugger or a way to get ddd to load and interpret a C source file and step through it a step at a time without requiring the debug-symbol compiled executable? I seem to remember doing something like this a long time ago with one of Borland's IDEs, but I might be mistaken. Well, it should be possible to create a script that does a quick compile of a given C program with debugging turned on and then starts gdb. This is what Borland IDE's (and Microsoft's, and in fact any IDE I've seen that allows one to debug) in fact do. However, you're really going to need an IDE to do that properly, as you need to tell the script how to compile your program (e.g. what -l flags to use). If you had a proper makefile, you could simply have a script that does make debug and then invokes gdb; if you can get ddd to use that instead of the real gdb, this may do what you want.
Re: RESCUE DISK | X ICONS
BOB'S MAIL [EMAIL PROTECTED] writes: 2. How do I install incons for applications in X -- or set a desk top etc.? You need to install a window manager other than the basic twm which comes with X. I'd recommend fvwm2 or fvwm95 for people just starting out. You can customize these relatively easily and they come with loads of icons.
Re: PON dials out, but PPP connection dropped
David Karlin [EMAIL PROTECTED] writes: Hello, I've been messing around with Debian/Linux for about 3 months now. I recently acquired a modem for my hamm box and am trying to set up my system for a ppp/pap connection. I have run pppconfig, and now when I do pon, the system waits about 1/2 minute, and dials the ISP's number. After the modem dials the number, I can hear the carrier-detect; then the modem disconnects in about another 1/2 minute. The same thing happens (more or less) when I run minicom. I have attached some of some related files. Also, this computer is on a LAN (in case that makes any difference). I got some help today (in IRC) from someone who works at an ISP that runs Debian, and he was stumped. If anyone can figure this out, I will owe you a beer (or six). TIA, --David The wait between the pon command and the modem dialing is telling - that wait shouldn't be happening. I'd suspect an IRQ problem - that is, your modem is set to use the wrong IRQ. When this happens, data comes back from the modem only very slowly; this means that there's a long wait for the OK in the second line of the chatscript, which explains the initial delay in dialing. This also interferes with data coming back from your provider in a timely fashion, which is leading pppd to think that a timeout occured in getting config. information, and so it hangs up. Can you use your modem through minicom? If so, is it ungodly slow? That's a better sign that there's an IRQ problem. To fix this, you need to find out what IRQ the physical modem is using (check the modem manual to find this out; if it's an external modem check your BIOS - should be IRQ 3 or 4, but some bioses let you set the IRQ on serial ports to 12). Then, you need to set the modem device to use this irq: setserial /dev/ttyS1 irq ?? (where for ?? use what you found; instead of ttyS1 use what's appropriate for your system) If this works, modify the file /etc/rc.boot/0setserial to make this irq change permanent.
Re: Voice Chatscript
David Stern [EMAIL PROTECTED] writes: Hello, I'd like my computer to call up one of those automated voice business information systems, then execute some transactions by entering my user account information and make some selections from the menu. I'll need to execute this task repetitively while I'm away from home. I tried writing a chatscript, but my modem hangs up, presumably because there's no handshake on the other end (I hear the automated voice on the other end after answering, then after about 10 seconds my modem hangs up). I see chat is intended for use with pppd, and I didn't see any options to disable the handshake expectation. ---chatscript--- ABORT BUSY ABORT NO DIALTONE ATDT777- (phone no.) \d\d\d\d\d\d\d\d\d\d (more than enough time to answer) \d\d\d\d\d99#\c (hear dialogue, enter user info followed by #) [..never gets past the initial dialogue, my end hangs up after 10 s..] As other people have pointed out, it's not chat that's hanging up, it's your modem. chat knows nothing of how long to wait for a handshake from the remote modem, etc. Besides that, even if your modem didn't hang up, this wouldn't work. What's the modem supposed to do when it sees the ? It's not dialing anymore - you finished that ATDT line. The modem will think you're trying to send the string 9 to a remote modem, or that 9 is supposed to be some modem command. In any case, it's not going to work. To make it work, don't depend on the delays in chat to separate key tones - let the modem do the delays in the dialing and just think that it has to dial a long number before it hears anything. So, your chatscript would look like: ABORT BUSY ABORT NO DIALTONE +++ ATZ OK ATDT777-,,,999#, \d\d\d\d\d\d\d\d\d\d\d\d\d\d\d\d\d The +++ and ATZ are to set the modem to its default configuration when we start all this. Then, the long phone number given to the modem is what to dial; commas mean delay (how long is configurable in one of those S registers) - on some modems, you can use @ signs instead of commas for longer delays - check your modem manual. The \d's at the end say how long chat waits after telling the modem to start dialing before chat sends a return character, which will hang up the modem. (so that after doing what it needs to do, the modem isn't still sitting around waiting for some modem on the other end; remember, there's no way to tell in a voice conversation if that click you just heard was the other person hanging up or just line noise without asking them, so the modem won't be able to tell that the other end hung up after getting the # sign) Adjusting the numbers of commas and \d's should get this to work.
Re: this is probably too easy
Are you certain that your chatscript is going far enough? That is, are you certain that after sending PPP your ISP doesn't ask for a username and password? (or do you perhaps need PPP in lowercase?) This error looks like your ISP isn't responding to pppd at all; this often happens when there's some other prompt your ISP presents that you aren't taking care of with your chatscript. Try dialing in manually with minicom, and do what's necessary to start ppp far enough to get the ppp garbage that looks like !}!}!} }8}}} } } } }#} or similar. (or, at the very least get to the point where your ISP isn't sending you anything, but is waiting for pppd on your machine to send garbage like the above). Then, to make sure that you have indeed manually connected properly, exit minicom without hanging up the modem (Ctrl-A Q), and do: /usr/sbin/pppd bsdcomp 15 crtscts mru 300 mtu 300 defaultroute noipdefault /dev/modem 38400 modem noauth (all on one line). (Instead of 38400, use the speed minicom was using with your modem; if your provider uses PAP, include user yourusername at the end of the line). Hopefully you'll find an error somewhere in your chatscript. Jeremy Freeman [EMAIL PROTECTED] writes: i have been pulling my hair our over this one and it is something so simple you will laugh at me My PPP doesnt work now dont all roll your eyes let me explain my chat script works and my pppd works but when ever i try and open a link to my isp it all terminates and puts this error in my messages file Sep 16 00:10:09 localhost chat[760]: Protocol: Sep 16 00:10:09 localhost chat[760]: -- got it Sep 16 00:10:09 localhost chat[760]: send (PPP^M) Sep 16 00:10:09 localhost pppd[755]: Serial connection established. Sep 16 00:10:11 localhost kernel: PPP: version 2.2.0 (dynamic channel allocation) Sep 16 00:10:11 localhost kernel: PPP Dynamic channel allocation code copyright 1995 Caldera, Inc. Sep 16 00:10:11 localhost kernel: PPP line discipline registered. Sep 16 00:10:19 localhost kernel: registered device ppp0 Sep 16 00:10:19 localhost pppd[755]: Using interface ppp0 Sep 16 00:10:19 localhost pppd[755]: Connect: ppp0 -- /dev/modem Sep 16 00:10:49 localhost pppd[755]: LCP: timeout sending Config-Requests Sep 16 00:10:49 localhost pppd[755]: Connection terminated. Sep 16 00:10:49 localhost pppd[755]: Receive serial link is not 8-bit clean: Sep 16 00:10:49 localhost pppd[755]: Problem: all had bit 7 set to 0 Sep 16 00:11:30 localhost pppd[755]: Exit. Sep 16 00:13:10 localhost kernel: PPP: ppp line discipline successfully unregistered
Re: printer advice
[EMAIL PROTECTED] (Greg Norris) writes: As near as I can tell it's not a winprinter, and it's print-language is Enhanced HP PCL 5 (which I believe is a superset of postscript, but might well be mistaken on that point). No; it's a superset of HP's PCL (Printer Control Language) - a proprietary HP language. (PCL is kind of interesting; it reminds me of the good old days of sending escape codes to my parent's first dot-matrix printer, and yet one can do vector graphics with it) But that's ok; ghostscript can convert postscript into the PCL that was used on HP LaserJet 4's, so a simple little filter in /etc/printcap will take care of being able to print postscript. (or you could just take care of things by installing magicfilter which ends up moving everything through ghostscript)
Re: Easy X exit manager....
Person, Rod [EMAIL PROTECTED] writes: Hey all, My wife has taken a liking to KDE's kpoker. I added her as a user, but she can't grasp the shutdown process. She is very computer impaired (I'm being nice...she just plain computer stupid.) Anyway is there a exit manager available to go from X to system halt state, or is this going to be my first attempt at linux programming. Someone posted at one point a simple little bit of tcl/tk code that added Reboot and Shutdown buttons to the xdm login screen. I'm not certain if that's exactly what you're asking for (it would require logging out of kwm and then hitting a shutdown button), but it's easier to remember than switching to a VT and then doing shutdown -h now. Hopefully someone on the list with a better-connected machine can find it in the archives.
Re: FTE editor
(I'm copying so much text because the original didn't make it to debian-user) Helge Hafting [EMAIL PROTECTED] writes: In [EMAIL PROTECTED], on 09/15/98 at 10:08 AM, Daniel Martin [EMAIL PROTECTED] said: Paul M. Foster [EMAIL PROTECTED] writes: [...] 2) Is there a liability to changing the permissions on these device files so that regular users have r/w access to them? Well, how comfortable are you with the ability of anyone logged in (or even with a process running) on your machine being able to grab the contents of any of the virtual consoles? If you do this, then anyone will be able to grab anything that appears on the screen. It's not as bad as xhost +, since they won't be able to send keys to, say, your root shell, but the ability to log everything may be a bit unnerving. Also, there's major nuisance potential since they could make any virtual screen display anything. I havent tested this yet, but consider the following: There is a file in /etc (sorry, don't remember which one) that can specify what groups a user will be added to when logging in on the console. One documented use for this is to grant membership to group audio so that anyone currently logged in on the console may use the audio device. Surely this trick could work with /dev/vcsa*, set the group to audio or create a new group for this purpose. Note that the audio trick isn't on by default, you must edit that file. (Do a grep audio /etc/* in order to find what file this is in.) The reason is that a hacker user is able to get permanent membership in the groups listed. Using this is still better than granting anybody access to /dev/vcsa as many users don't know the hack involved, and I believe they need to use the console in order to do it. No problem if the hacker never get near the console. True; (the file is /etc/login.defs). However, I'd not call the way one gets access to one of these groups permanently a hack - I'd call it basic Unix knowledge. (I mean, if you know what it means to have a program setgid and know how to make a program setgid, you've got it). But yes, if the console is in a secure environment, then there's no risk in doing this.
Re: X-modes other than the defaults.
Mrpeabody [EMAIL PROTECTED] writes: My X-windows automatically loads in 8dpp mode and I have a card that is capable of much better how do I tell x to try other modes instead. -jeff Two ways: If you're starting X with startx, you can do: startx -- -bpp 16 (or 15, 24, or 32) If you're starting X with xdm, you need to edit the /etc/X11/XF86Config file to change the default bpp. In the Screen section put: DefaultColorDepth 16 (or whatever number). This will also affect the default bpp used with startx. You'll have to log out of any xsession xdm is managing; if your xdm is at the login screen, press Ctrl-Atl-Backspace to kill the X server; when it restarts it'll be using the new bpp.
Re: Debian 2.0: pon works, diald gets PAP authentication failure
Ken Westerback [EMAIL PROTECTED] writes: I have a working ppp connection, using PAP, available through pon or wmppp as evidenced by the console messages: SNIP When I try to enable diald by filling out the diald.options files as follows: fifo /etc/diald/diald.ctl mode ppp connect /usr/sbin/chat -f /etc/chatscripts/provider device /dev/ttyS1 speed 115200 modem lock crtscts local 127.0.0.2 remote 127.0.0.3 dynamic defaultroute pppd-options asyncmap 0 include /etc/diald/standard.filter and running diald from the command line I get a PAP authentication error: SNIP Can anyone tell me what the problem might be? It appears to me that pppd may not know what username to use for PAP authentication. Try using: pppd-options asyncmap 0 user YourPPPUserName Now, one effect of this is that YourPPPUserName (but not password) will end up in the logs. If this is undesirable, you could create a file called /etc/ppp/peers/dialdopts and in that file place: asyncmap 0 user YourPPPUserName And then in /etc/diald/diald.options just have: pppd-options call dialdopts
Re: FTE editor
Paul M. Foster [EMAIL PROTECTED] writes: Debs: SNIP Two questions: 1) Is this a dangerous thing (I know of no other Linux editors that access the terminal this way)? Accessing the terminal in this way? In itself no; in fact, if all access is done through /dev/vcsa*, there's not the chance of leaving your console in the lowercase characters appear as line-drawing characters mode. However, using an untested program as root is just one of those generally unadvised actions. 2) Is there a liability to changing the permissions on these device files so that regular users have r/w access to them? Well, how comfortable are you with the ability of anyone logged in (or even with a process running) on your machine being able to grab the contents of any of the virtual consoles? If you do this, then anyone will be able to grab anything that appears on the screen. It's not as bad as xhost +, since they won't be able to send keys to, say, your root shell, but the ability to log everything may be a bit unnerving. Also, there's major nuisance potential since they could make any virtual screen display anything.
Re: FTE editor
Paul M. Foster [EMAIL PROTECTED] writes: In my case, it isn't a problem, since it is a single user machine on a network, with my wife running Win95 on the other machine. But I see the point. So presumably, if someone could write something to /dev/vcsa*, it would show up on my screen. Hm. Up until I unpacked this editor, I hadn't even heard of /dev/vcsa*. /dev/vcsa* and /dev/vcs* devices are nice, even if using them does make it a bit more difficult to port FTE to screens that aren't virtual consoles (like xterms). It's much better that a program use these devices than try to open /dev/kmem (Yes, someone wrote an editor ostensibly for linux that did that. An evil, evil program.) It would be nice if login could set perms and ownership on vcs* devices the way it does on ttys - that would be the ideal solution. Try this: As root (or any other user who can access /dev/vcs0), do: cat /dev/vcs0 /tmp/storevcs then do something which does stuff to the screen - ls or the like. Then: cat /tmp/storevcs /dev/vcs0 (vcs0 means whatever vc is currently displayed). And instantly your screen image will go back to what it was at the first cat. This is an example of what people could do to your screen.
Re: PON on REQUEST
Jim Crumley [EMAIL PROTECTED] writes: I pretty much copied my setup from http://www.math.jhu.edu/~martind/mybox.html . I think Daniel Martin still reads this list. He also has a link to the Dynamic IP Hacks Mini-HOWTO, which has more suggestions. Oh my. I'd hate to think I'm an authority. I have something like this set up - I have a voice modem, and use vgetty (in the mgetty-voice package). I've modified some of the scripts that came with it (but not much; they were pretty trivial modifications) so that I can call my machine from anywhere, enter *, a secret code, and # and my machine will dial in. (A different code lets me check my messages - my voice modem is my answering machine) That's not documented on the page mentioned above, (I can send you details, but using xringd may well be easier) but how to have a page automatically updated is; although, there's a way that I think is better, which I used on my girlfriend's machine since her ISP didn't allow her shell access to set up a .forward file. Basically, in your /etc/ppp/ip-up script (for bo, aka Debian 1.3.x) or in some file in /etc/ppp/ip-up.d/ (for hamm, aka Debian 2.0 - you'd probably be best off putting this in a new file, something like /etc/ppp/ip-up.d/local_updatepage) put the following: m4 -Dipaddr=$PPP_LOCAL /tmp/newip.html EOD HTMLHEADNew ip address ipaddr/HEAD BODYH1New address is: ipaddr/H1 PNew address ipaddr obtained on syscmd(date)/P/BODY/HTML EOD HOME=/root ftp my.webspace.provider EOD send /tmp/newip.html public_html/newip.html EOD (In bo, use $4 instead of $PPP_LOCAL in the top line) Then, in /root you need to create a .netrc file giving your username and password for whatever machine you're putting this webpage on. The syntax is: machine my.webspace.provider login myusername password mypassword Then, remember to set all the permissions correctly: chmod 0600 /root/.netrc chmod a+rx /etc/ppp/ip-up.d/local_updatepage # or whatever # Remember, use /etc/ppp/ip-up on bo.
Re: diald and ppp
[EMAIL PROTECTED] (Keith) writes: [1 text/plain; US-ASCII (7bit)] I am trying to set up diald and ppp. I thought it would be easy like it was with Debian 1.3, but I guess Debian developers or someone wanted to keep me on my toes so they changed stuff. Here are my problems. SNIP I have attached /etc/diald/diald.options and /etc/diald/connect I am using pap-secrets also. # /etc/diald/diald.options SNIP connect /etc/diald/connect disconnect /etc/ppp/ppp-disconnect SNIP [3 /home/nfn11988/connect text/plain; US-ASCII (7bit)] TIMEOUT 5 ATZC1D2%C0 OK ATDT4038000 ABORT NO CARRIER ABORT BUSY ABORT NO DIALTONE ABORT WAITING TIMEOUT 45 CONNECT TIMEOUT 5 The problem seems to be that for diald's connect script you're just giving it a chatscript; diald doesn't want that, it wants an actual script. Try moving your chatscript to another file (say, something in /etc/chatscripts) and then in your /etc/diald/connect file just have: #!/bin/sh /usr/sbin/chat -v -f /etc/chatscripts/whateveryoumoveditto Then, don't forget to make /etc/diald/connect executable.
Re: analog 3.0.3 home grown scripts
William R. McDonough [EMAIL PROTECTED] writes: could someone please show me where to access these home grown scripts I have analog on my web server but it's FreeBSD not Debian... I'd love to have my analog show me email and FTP stats. ftp://ftp.debian.org/debian/dists/slink/main/source/web is the directory you're looking for. Incidentally, it's 3.0-3, not 3.0.3 - the difference is that Debian uses a dash to indicate the difference between packaging versions and upstream versions. 3.0-3 is the third Debian package built out of analog 3.0, not an improvement to analog 3.0 put out by the analog authors. The file you probably want from there is analog_3.0-3.diff.gz which shows you the new stuff debian added to the source of analog; you can read through the diff and figure out what files you're after. Alternatively, if you don't feel like sorting the debian specific stuff out of the .diff.gz file, you can download the analog .deb and just unpack it on your system. That's right, you can unpack Debian .deb files using only standard Unix tools. (One of the arguments people sometimes give about why .deb format is superior to .rpm format) You just: cd /tmp mkdir analogdebian cd analogdebian ar x /path/to/analog_3.0-3.deb gunzip data.tar.gz tar xvf data.tar And analog is now unpacked for you in subdirectories of /tmp/analogdebian. The example scripts are in usr/doc/analog/examples. The .deb file is available in: ftp://ftp.debian.org/debian/dists/slink/main/binary-i386/web
Re: changed server name, how-to change back
[EMAIL PROTECTED] (Keith) writes: I somehow changed the name of my server back copying files from another server onto this one. Which files sets the name of the server? It took me a moment to figure out what you were asking. The file you're looking for is /etc/hostname, which is used in the startup process to set the machine's name.
Re: ppp connection with Demon
Tom Bowles [EMAIL PROTECTED] writes: Incidentally it's not just receiving and sending data that is slow. There is also a 20-30 second delay between sending the modem initialisation (and phone number for that matter) strings and getting any response. Could it be something to do with the serial interface itself. I am now about 100 miles from anything I know anything about, but it's a hell of a lot more interesting than windows. I think you may be onto something here - since you're getting data, but there are serious timing delays (that you're certain aren't caused by Demon being slow - that is, delays even in the init strings), you may have the wrong IRQ set on your serial device, or there may be some conflict in IRQ settings. Do a: setserial /dev/modem (or /dev/ttyS1, or /dev/WhateverYourModemIs) and note the results. Since you are getting data, the port setting is correct and the UART listed is correct or close enough. The IRQ, however, may well be wrong. If you have some other OS in which the modem works well, you could look at the device information for the modem there. Once you have the correct IRQ for your modem, you'll want to edit /etc/rc.boot/0setserial down around lines 25-30 and fix the information there. You can then as root run /etc/rc.boot/0setserial and try your modem again. Is the modem a plug-and-play internal modem? If so, look at using the isapnptools to set up the modem. (You'll have to run one program to dump out the configuration info. on your modem, edit that file to suit your needs, and then run another program on the edited file to set the configuration in your modem - once you get the config. the way you want it, putting it into /etc/isapnp.conf will have it used at boottime - but there's much more info. in the isapnp docs) Note that if you have an external modem that calls itself plug-and-play, this doesn't actually mean anything. (Well, not anything that you care about with regards to setting up your modem for linux) If the modem is an internal one with switches/jumpers, you'll need to look at it and your modem manual to determine what IRQ it's set to. (If you've lost the modem manual, you might want to check the manufacturer's web site). If it's an external modem, then the IRQ's are probably 4 for /dev/ttyS0 (aka com1) and 3 for /dev/ttyS1 (aka com2). However, some machines (like mine) have BIOSes which can alter this, so that the built-in com ports use IRQ's 4 and 12. Check your BIOS's setup menu. DANIEL MARTIN p.s. What are all these settings: IRQ, io port (sometimes called just port) and the UART number? Well, UART is just a number that identifies the type of chip controlling the serial device. The kernel needs to know this so that it can know how large a buffer the device has, what speeds it supports, how it reports line conditions (like hardware flow control), etc. UART type can usually be detected automatically. The most common UARTs are the 16550, 16550A, and 16650. This number is something specific only to serial devices. The io port and IRQ require a very brief explanation of how the processor gets data from other devices in the system. Essentially, all devices in the system are connected to one set of wires called the data bus. The processor, when it wishes to send or get data, will set some wires to identify the address it wishes to access, a wire to indicate whether this is a read or a write, and will then either set the relevant data wires for a write or assume that whatever device responds to this address has set their state as soon as it requested the read and will then read the data wires. (this is oversimplified, but you get the idea). Anyway, the io port is the address the processor uses to get and send data to the modem. (actually, serial devices generally need 8 consecutive address; you specify only the first one - this is why this will sometimes be called the base io port address or io base; these ports are used not only to send and receive data but also to set/get status information) You can see what devices use which io ports on your system by doing cat /proc/ioports. Now, when a device needs to tell the processor something (like incoming data or my output buffer is empty, you could send more), the device can't force the processor to do a read. Instead, what the device does is raise a signal saying in essence pay attention to me. It does this through an Interrupt ReQuest, or IRQ. There are, I believe, only 16 of these lines available and several are allocated to system things (like the separate chip that controls the clock, or the keyboard, or the internal line used to signal floating point math errors). With an IRQ set incorrectly, the processor would not realize that there was incoming data on the modem - this could cause the unbelievably slow response time you seem to have. You can find out which devices use which irq lines by doing cat /proc/interrupts.
Re: resolv.conf, PPP and multiple providers
(This is being resent because it was mercilessly bounced... :( ) [EMAIL PROTECTED] writes: *- Damon Muller wrote about resolv.conf, PPP and multiple providers | | However, anyone have any hints on how to set up the resolv.conf file | when you have multiple ISPs? I guess I could have an ip-up script modify | it when it dials each ISP, but that's a bit of a pain. | That is about the only way to do it. Read /usr/doc/HOWTO/mini/Dynamic-IP-Hacks.gz. It is a little out of date but it has some good ideas. Another file to modify is /etc/hosts. Note that under Debian 2.0, /etc/ppp/ip-up is a script that runs every file under /etc/ppp/ip-up.d/ - so if you want to add something it's easiest to do it by adding a new file to that directory (note that the filename must only contain the alphanumeric characters and underscore(_) ). On simple way to do this is: #!/bin/bash get_ns () { case $PPP_REMOTE in 128.220.222.2) echo 128.220.2.7 128.220.2.82 ;; 137.22.*) echo 137.22.1.13 137.22.1.15 137.22.1.4 ;; 203.20.*) echo 203.20.80.1 203.34.5.3 ;; esac } echo -n /etc/resolv.conf # Next line optional echo search empire.net.au /etc/resolv.conf for ns in `get_ns`; do echo nameserver $ns /etc/resolv.conf; done
Re: xdm starts local server
(This is being resent, as I have had to reconfigure my mail program to get around internic's objection to the debian.org name) Hamish Moffatt [EMAIL PROTECTED] writes: On Aug 12, 1998 at 03:20:30PM +0200, [EMAIL PROTECTED] wrote: Why you get different behaviour from xdm and start might be explained by this: There's a script /etc/X11/Xsession that claims to be run by both xdm and xinit (to which startx is a wrapper.) The script seems to look for ~/.xsession only though. Maybe it isn't run after all from xinit on Debian systems. That means that the comments in the file are misleading. I don't actually have either ~/.xsession or ~/.xinitrc, which is why I find this all the more strange. Ok - what about your /etc/X11/xinit/xinitrc file? This should be a symbolic link to ../Xsession; if it isn't, then I can understand why you get this odd behavior. If it is a symbolic link, then something must be going wrong in the /etc/X11/Xsession script when started by startx. Tell me, is the file ~/.xsession-errors created when you start X with startx? (Delete the file first, as it's surely being created when you log in via xdm) If so, do the contents of that file after starting X with startx provide any clue as to what's going on?
Re: fvwm95 after hamm upgrade
Doug Thistlethwaite [EMAIL PROTECTED] writes: Hello, I upgraded my bo system to hamm last night. Everything went pretty well, though I did have the run the install portion of dselect several times for it to get through without errors... Well, I'm impressed - I made a slight error in packaging up fvwm95 for hamm, such that bo upgrades aren't as smooth as they could be - oops. Anyhow - Everything seems to have installed fine. But I have now noticed that my windows manager fvwm95 does not seem to read my prior config file in my home directory The config info was in /home/doug/.fvwm2rc95 on my bo system and everything seemed to work fine. The first lines of my window-managers /usr/X11R6/bin/fvwm95 /usr/X11R6/bin/fvwm2 ... What do I need to do so my windows manager will read my ini file? The name of the ini file changed between the bo and hamm versions - if you had a customized system-wide initialization file, you got a little message about it. Personal initialization files are harder to check for. Anyway: 1) The name of the new initialization file is .fvwm95rc - you can rename your .fvwm2rc95 to .fvwm95rc and it'll work, BUT 2) fvwm95 now has an improved configuration, much like the configuration scheme of fvwm2 - if you use this new scheme then whenever you install new packages, they'll appear in your fvwm95 menu the next time you restart fvwm95. For more information on this new scheme, refer to /usr/doc/fvwm95/README.sysrc.gz. The basic gist is to create a subdirectory called .fvwm95 and then place in it a file called post.hook containing your own customizations. There are also some examples of things you can do with the system configuration files in the system config. files themselves (they're all commented out; uncomment what you want). Note that if you do have a ~/.fvwm95rc file, the new configuration scheme won't get touched, and your menu won't get updated to show the list of currently installed packages. Some people who are used to the one long monolithic configuration file find the new config. scheme a bit confusing - but it's not, really. I'd be glad to help you with it.
Re: minimal files essential for booting ?
Brian May [EMAIL PROTECTED] writes: In article [EMAIL PROTECTED] you write: I went nuts partitioning the new disk. I was mostly just experimenting. Many will think I have gone needlessly overboard. I won't disagree. The old disk has two partitions, one being swap. The new disk has a partition for just about everything. These directories all live on their own partitions: /usr /usr/local /var /home /etc /bin /tmp /lib I think the FSTND standard requires these directories to be available during boot: /bin /dev /etc /lib /tmp /sbin (Note that /etc and /tmp must be writable, I don't know about the rest.) I can't remember about /var, but suspect that it is required for /var/run, /var/lock, and /var/log. /var can be on its own partition without any problem - in fact, it is possible to have /tmp on its own partition too, but in that case there must still be some room in the / partition for a small /tmp directory that the system can use before mounting the real /tmp. The rest of your partitioning setup, though, seems fine. Just move /lib, /bin, and /etc back to your / partition and all should be fine. And the boot process (and going into single user mode for maintenance/backups) is the only time when you have to worry whether enough programs are present on certain partitions. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: xacc brings my laptop to a crawl
Mike Roberts [EMAIL PROTECTED] writes: I'm trying to use xacc in the hamm release (xacc_1.0.17-2.deb), but when I try to enter a number in the register window, the cursor blinks exceedingly fast and it brings to X server to a near standstill. I can still continue, but at a snail's pace, maybe a character every 10 seconds. This occurs with both the lesstif and statically linked motif version. So, does xacc have a bad polling mechanism or something or do I have my Xserver misconfigured? Mike Roberts This is actually a bug in lesstif. (More accurately, in the libc6 lesstif contained in the package called lesstifg) You can fix it (supposedly, according to the lesstifg maintainer) by installing the lesstifg from slink, version 0.84-1 or higher. (I think the current version is something like 0.85.3-1) Once you do install this, I _strongly_ suggest putting lesstifg on hold (with an equals sign) in dselect. Unfortunate mistakes with version numbering means that the hamm lesstifg will try to come back if you don't do this. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: Diald and Windows NT/95
Butch Kemper [EMAIL PROTECTED] writes: 1. What is causing the link to yo-yo and how to change the Windows boxes to stop the yo-yo from happening. Doesn't diald come with some sort of monitor program? (use 'dpkg -L diald' to see what comes with diald) I think it's called 'dialdmon' or something like that? Why don't you watch that to find out what the windows boxes are doing? -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: SHELL: duping stdout and stderr to another file
Ulisses Alonso [EMAIL PROTECTED] writes: Hi all I would like to know _how to copy_ (not start a new shell, eg: script) stdout and stderr _from a shell script_ to a file. Also It is interesting for me if there is a way to stop copying stdout and stderr... Thanks in advance, Ulisses Well, I'm not entirely certain what you mean, so I'll just say several things that might answer your question and hope that one of them tells you what you want to know. Suppose I have a simple shell script: #!/bin/sh echo Guvf vf n frperg zrffntr. | tr a-zA-Z n-za-mN-ZA-M And I call this script foo. I make it executable so that I can call it like this: cush:~$ ./foo This is a secret message. Now, I decide I want to redirect this into a file. I do: cush:~$ ./foo sekritfile Ok. So I can redirect all of foo's stdout into a file. But suppose I made a mistake in foo (say, I added an extra argument to the `tr' command). Then I get: cush:~$ ./foo sekritfile tr: too many arguments Try `tr --help' for more information. Well, I want to redirect both foo's stdout and stderr into a file. Ok; I do this by saying redirect stdout to sekritfile and redirect stderr to the same place as stdout: cush:~$ ./foo sekritfile 21 All well and good. I can even do fancy things like redirecting stdout and stderr to two different files: cush:~$ ./foo sekritfile 2 errfile Ok. So far so good. Now suppose I don't want to have to specify where the stuff in foo goes on the commandline; suppose I want the shell script to just redirect everything all the time. I can do this by using the shell builtin command 'exec' in a slightly unusual way: #!/bin/sh exec sekritfile 2 errfile echo Guvf vf n frperg zrffntr. | tr a-zA-Z n-za-mN-ZA-M So now what happens is that when I do: cush:~$ ./foo Is that the output of the tr command is redicted to 'sekritfile' and the stderr of the tr command (and any following commands) is redirected to 'errfile'. I could redirect them both to the same place using: exec sekritfile 21 So far so good. But what if I want to redirect something and then undo the redirection? Say, if I had: #!/bin/sh echo Guvf vf n frperg zrffntr. | tr a-zA-Z n-za-mN-ZA-M echo Done decoding. And I wanted the first line redirected, and the second not. Then, I just need to save the original destinations of stdout and stderr, like this: #!/bin/sh exec 31 42 sekritfile 21 echo Guvf vf n frperg zrffntr. | tr a-zA-Z n-za-mN-ZA-M exec 3 24 echo Done decoding. Then, only the first echo command gets redirected to a file. cush:~$ ./foo Done decoding cush:~$ cat sekritfile This is a secret message. But notice: cush:~$ ./foo msgfile cush:~$ cat msgfile Done decoding. cush:~$ cat sekritfile This is a secret message. Now, suppose that instead of redirecting output, I want to copy it. (That is, I still want to see the output on the screen). This means the output has to go to two places - the file and the screen. There are a few ways to do this - I'm only going to show one involving the 'tail' command. The problem is that the shell won't split output by itself, and won't do overall redirections (like the exec commands above) into pipes. However, we can do this: #!/bin/sh cat /dev/null sekritfile # create it so that tail doesn't complain tail -f sekritfile tailpid=$! exec sekritfile exec 21 ... ... more commands here ... sleep 2 kill -1 $tailpid This causes the shell to dump stdout and stderr into the file sekritfile and at the same time starts up a process (the tail -f) that reads the file and displays its contents. The last kill command is necessary to stop the 'tail' process when the script is done. If you want to do this and only want to redirect portions of the script, you can do: #!/bin/sh cat /dev/null sekritfile tail -f sekritfile tailpid=$! exec 31 42 exec sekritfile exec 21 ... ... these commands get saved to the file ... exec 51 62 13 24 ... ... these commands don't go to the file ... exec 15 26 ... ... these commands go to the file ... sleep 2 kill -1 $tailpid I hope this answers your question. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: no pon!
Martin Bialasinski [EMAIL PROTECTED] writes: DP == David Parmet [EMAIL PROTECTED] writes: DP who is this pid 109 and what does it want with my life? pid = Process ID To check what prozess has ID 109 do a ps ax|grep 109 Incidentally, if you have the process id already, all you have to do is: ps pid That is, ps 109 The advantage of this is that you also get the headers telling you what each column means; for example when one does: ps u 1 And gets: USER PID %CPU %MEM SIZE RSS TTY STAT START TIME COMMAND root 1 0.0 0.2 76876 ? S Jul 26 0:03 init [2] One has at least some chance of figuring out what all that information means. The ps ax | grep foo idiom is so common that it's sometimes easy to forget that ps does take non-option arguments. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: partition memory
Yanick MICHOU [EMAIL PROTECTED] writes: Hello, I would like to know what is the shell command to know what is the available, used and free memory on the current partition ? You mean the available used and free disk space? The command is df - I believe it stands for something like disk free. It'll tell you how much is available on each partion. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: Fvwm2 and Debian menu
Curt Daugaard [EMAIL PROTECTED] writes: I like the design of the Debian menu package, and I'm trying to follow the docs to get auto-updating of menus under X, but I seem to have hit a roadblock. I use Fvwm2, and according to the comments in system.fvwm2rc and the Fvwm2 docs--in the Debian package, compliant with the menu package--FvwmButtons (formerly Good Stuff) is set up so that the user can easily override the defaults. But it looks like the menudefs.hook, autogenerated by the menu-package, is determined to invoke DebianFvwmButtons, which stubbornly overrides any configuration the user attempts. Well, you don't have to start your FvwmButtons from the menu - you can just start it automatically (assuming you always want it to be there); on one account on my machine, this is done by putting all the customization into post.hook and a call to 'Module FvwmButtons' in init-restart.hook. If, however, you want it in the menu you can always put it in yourself with an AddToMenu command in post.hook; modifying the menu entry itself (by changing /usr/lib/menu/fvwm2) gets wiped by a new fvwm2. This has the problem that the DebianFvwmButtons is still there... The best solution is probably to use DestroyModuleConfig DebianFvwmButtons in post.hook and then listing the entries one wants. Of course, since this command exists, one wonders why the author of the fvwm2 config used DebianFvwmButtons to begin with... -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: No PINE debian package?
jason and jill [EMAIL PROTECTED] writes: It's included in the distribution and available for download from www.debian.org. Can't get much more available than that unless you contract Pam Anderson to deliver it to your home. ;) He may have missed it because it's in non-free rather than mail. Also, note that Pine is not available directly as a binary for hamm; it was actually a mistake to distribute it for bo (Debian 1.3.x) in the first place, and the University of Washington has asked us not to distribute any precompiled binary other than the ones they've approved; however, we can distribute the source, and so in hamm pine exists only in source form. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: Why does tty1 become the current VT on reboot/halt?
Noah L. Meyerhans [EMAIL PROTECTED] writes: Hey all. Why is it that when I reboot or halt my machine, I'm automatically switched to the first virtual console? I don't like this at all, because all of the messages coming from the rc scripts get sent whichever VC I was on when I typed the reboot/halt command. I haven't been able to figure out how/where the event occurs. Is it possible to prevent it, or to have all the rc scripts send their output to tty1? Well, I can't provide a solution, but I can provide an explanation: xdm, or more specifically the Xserver xdm was running. Whenever an xserver shuts down, it reverts to the virtual terminal that was active when it started (which is usually VT1, unless you hit Alt-F# at the right time in the boot sequence). While this makes sense as something to do when the xserver is shut down directly, (for example, if one were running X through startx and exited X, one would expect to go back to the shell prompt that ran startx) it may not be the right thing to do when the Xserver is shut down while not active. I don't find this any more than a really minor inconvenience, since switching vt's with Alt-Fn works during the shutdown process, and even continues to work after the system has halted. You might file a wishlist bug against your favorite xserver though saying that it shouldn't switch vt's on shutdown unless it (the Xserver) was on the currently active vt. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: My 'h' key is unusable
Stelios Parnassidis [EMAIL PROTECTED] writes: Just had the base system installation of Hamm, and wanted to type 'which superformat'. 'h' beeps and beeps ... i have to type Ctrl-V h to use it. I looked into the /etc/terminfo/l/linux via ... (untic) could find nothing. What's wrong with my very 'h'? Heh? :( Could it be a missing backslash or caret in your /etc/inputrc? A mis-quoted section of your bash initialization files? It sounds almost as if bash is set to interpret h as delete-next-character (something that people sometimes want the delete key to do; I can see how on a poorly thought-out installation, one might want Control-h to do this). See if h does indeed behave this way by typing some stuff, backing up, and seeing if you can delete things with h. To track down the problem, I'd suggest doing: (at the bash prompt) bind -p | less and inside less search for h (less doesn't use readline, so it shouldn't be affected by the weird-h stuff). Then, I guess I'd look at /etc/inputrc, ~/.inputrc, and the various bash initialization files for anything that might be causing this. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: FVWM2 (New X user)
Matt Kopishke [EMAIL PROTECTED] writes: I have X (finally) working on my machine (bo), I am running FVWM2 and was wondering what I do to add a pattern or color (prefer pattern) to the desktop. I now you edit the fvwm.rc file, but what do I edit, what format file do I use for the pattern, and were would I put it? Actually, the debian fvwm2 setup is made for easy customization, so that you don't need to edit huge monolithic .fvwm2rc files. Just create a subdir. of your home directory call .fvwm2 - then put the background you want into one of the following files: ~/.fvwm2/background.xpm ~/.fvwm2/background.gif ~/.fvwm2/background.jpg (If multiple files are present, .xpm takes precedence over .gif which takes precedence over .jpg) You can also just change the color by putting the color name (like red or blue) into the file ~/.fvwm2/background.color Finally, you can get it to choose a random background by creating a file called ~/.fvwm2/background.list, and listing one background file per line. DANIEL MARTIN P.S. One mistake people often make is to download a ~/.fvwm2rc file from the net or use a friend's - in that case, you lose all the nice debian-specific stuff in Debian's fvwm2 setup. If you ever want to do something to tweak fvwm2's setup, just add lines to the file ~/.fvwm2/post.hook as though you were adding lines to your ~/.fvwm2rc. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: Xconsole vs security
Shaleh [EMAIL PROTECTED] writes: If your password is in pap-secrets it is supposed to remain just that -- a secret. I never see my password come wizzing by on xconsole and I use pap too. So, now we need to figure out why. The \q only works in a chatscript. Do you perhaps have the debug option uncommented in /etc/ppp/options? Or is there a debug in /etc/ppp/peers/provider? That's the only thing in the ppp sources even looks like it could cause a password to appear in a log. (Though admittedly, I haven't examined the sources in too much detail). Tell me, what does the message with the ISP password look like? -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: -post_data
[EMAIL PROTECTED] (Gerald V. Livingston lI) writes: This isn't debian specific, but what the hey. How does one find out the format needed to use the -post_data switch in LYNX to feed data to a remote CGI using a script? The data format is exactly what's sent to the server, so any reference on http should tell you. Briefly, it's the syntax used in URL's after the ? - that is, data consists of one or more key/value pairs separated by ampersands (). The key and the value are separated by an equals sign (=), and sometimes a pair may consist of only a key, with no equals sign or value. Within the key and value, characters other than [A-Za-z0-9] are encoded (*) - spaces (character 32) are replaced by plus signs (+) and other characters are replaced by %XX, where XX is the hexidecimal character code (note that spaces then can also be written as %20 - in fact, any character _can_ be written as a %-sequence, but only certain characters need to be written that way) I'm certain that some perl modules in libwww-perl already handle this format as well as getting info. from the server - you're probably much better off using them than trying to pass this to lynx, unless for some reason you're avoiding perl. * Actually, you're allowed to stick more than just the digits and the numbers in without %-encoding them. In theory, the only characters that need to be encoded are Control characters, space, double quote () and any of the characters in: ;/?:@=+#%, but many people encode anything except the letters and digits -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: Fvwm95: FvwmTaskBarAutoHide problems
Benno Overeinder [EMAIL PROTECTED] writes: Hi, I am using the Fvwm95 window manager and has enabled the infamous FvwmTaskBarAutoHide option. With Debian 1.3, the Fvwm95 taskbar appeared automagically if the mouse hits the bottom of a virtual desktop (i.e., generated a certain event which made the taskbar to be raised). However, if the virtual desktop is one of the lower most of the desktops (i.e, with DeskTopSize 3x2, the lower three), the specific event is not generated and the taskbar did not appear. With the Fvwm95 distribition that comes with Debian 2.0, this problem is solved, that is, in the lower virtual desktops the taskbar appears without any problems. But now the taskbar appears in the other virtual desktops very slowly and in steps. The mouse changes from shape each time the taskbar raises with a step, as if a series of X events are necessary to raise the taskbar instead of a single event. 1. I've seen the problem (and should fix it, since I'm the Debian maintainer for fvwm95) 2. The taskbar is being raised in a series of XMoveWindow calls, by two pixels each call. I can try to make that an option in the next Debian fvwm95 package; howver, there's much to do with getting Debian's versioning back in sync with the upstream source, among other things, so this may wait. 3. I don't know of a simple solution (I rarely use the TaskBarAutoHide myself.) I'll see if I can disable some X events on the TaskBar as it's moving, but this could cause other problems. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: libc6 upgrade crash my machine!!!!!
[EMAIL PROTECTED] writes: After running dselect and trying to upgrade all my old hamm packages including libc6, nothing works now it seems the libc6 package is corrupted or something, i've tried to mount the file system and run dpkg -i --root=/mnt --admindir=/mnt/var/lib libc6xxx.deb but it complains about preinstall and postinstall scripts ( -- not found --). Please Help There was very recently a libc6 package (2.0.7r-3) that died if you had an /etc/ld.so.preload file or the environment variable LD_PRELOAD set. I'd recommend temporarily moving that file (after mounting your drive with the rescue disk as above) if it exists and commenting out any definitions of LD_PRELOAD in /etc/profile or similar files. You should then be able to boot your system and upgrade to the latest (-5) libc6 that fixes these problems. (This should be on the ftp mirrors now; though I can't get through to ftp.debian.org at the moment, I can see it on non-us.debian.org, so it's probably on all the major ftp mirrors) You can then move /etc/ld.so.preload back and re-enable any LD_PRELOAD settings. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: Xconsole vs security
Nuno Carvalho [EMAIL PROTECTED] writes: Hi, I'm using Debian 1.3.1 and KDE Beta4. When I call the xconsole program I could almost activity on my machine but I think there's something wrong ... Sometimes on xconsole I could see my login and password as when I write them ! It's rigth !?!? I don't think so ! As I work as root and have a username on my machine there's no problem but if I add a new account if someone call xconsole could see my password to my ISP ! The question, I think, is that you are concerned because when you dial up, the password to your isp gets logged by the chat program, and so appears in the xconsole window. You worry that anyone you give an account to can call up xconsole and thereby see your ISP password, which would be a bad thing. Ok, to begin with you can make it so that chat doesn't log your password by putting a \q in front of it. In my chatscript (/etc/ppp.chatscript on a Debian 1.3.1 machine) I have: ABORTBUSY ABORTNO CARRIER ABORTVOICE ABORTNO DIALTONE ATDT4103660015 name MyISPlogin word \qMyISPpasswd This will replace your ISP password with all question marks (like: ?) in the logged messages. (This next bit is directed at the list) I was going to add more, but then I noticed that the pipe xconsole reads is world-read - does this strike anyone else as a security hole? Surely the information dumped into /dev/xconsole is as sensitive as that dumped into /var/log/messages, right? -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: problem with libpam0 and libpam-util in stable
Chris R. Martin [EMAIL PROTECTED] writes: I'm probably missing something obvious here, but I can't seem to get libpam0 and libpam-util to install, because aparently they depend on each other. Do I need to use dpkg --force ?? Thanks, Chris Try specifying both on the same dpkg command line: dpkg -i libpam0_whatever.deb libpam-util_whatever.deb Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null Does this message strike anyone as significantly less helpful than the former mail a message with the subject unsubscribe to [EMAIL PROTECTED] message? I'll grant you that people screwed that up too, but this seems to be asking for people to get confused. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: [Debian] Urgent: how to create Debian CD from non-debian UNIX system
Nico De Ranter [EMAIL PROTECTED] writes: Hi, we have a local mirror of the debian distribution that we use to install our Debian PC's. However next week we will have to install Debian on a PC which has no access to our network so we want to create a CD starting from the current mirror. unfortunately the mirror is located on a SUN and none of my Linux machines has enough harddisk to contain a CD. I tried installing the sources for debian-cd but they don't seem to contain any documentation. -( Thanks in advance, Nico Well, it is also possible to use CD images that other people have created (assuming that your SUN has good network connectivity and that it has hard drive space to spare) - if you want to make a CD for installing a bo (aka 1.3 or 1.3.1) Debian system you can get the official CD image in any number of places - the www.debian.org site should tell you where in detail. If instead you want a CD that can create a hamm (aka 2.0, or at this point really 2.0-beta, since officially 2.0 doesn't exist yet) system, you can look at the webpage http://www.uk.debian.org/debian-cd/cd-images/ , where Phillip Hands has put his 2.0-beta CD images (I'd suggest only using the 2.0beta1 images on that site, however). I know this isn't exactly what you asked about, but it may be much easier to go get someone else's CD image than trying to nudge the debian-cd scripts into working on a SUN. DANIEL MARTIN -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: Is Netscape Navigator really only for 16-bit color?
Chip Grandits [EMAIL PROTECTED] writes: I've installed Netscape Communicator v4.05 on my bo system running Xwindows with the FVWM manager. I don't have any Motif or Motif clones. When I attempt to browse to a page which uses Java, Netscape just crashes immediately and goes away. (No harm to anything else - as far as I can tell). The one thing I turned up from my own research is to ensure that Moz_2.zip is in a particular directory ( I think /usr/local/netscape - whatever it said that's where it is on my system). Does anyone know why it's going down? Do I need some additional Java Support Packages besides that which comes with Netscape? Do I need another archive utility for moz_2.zip. Is there a particular log file I can look at to get more info? Tell me - how did you install netscape? Did you install it with the Debian netscape-installer package, or did you attempt to do your own installation? If you installed netscape without using the Debian package, then try starting netscape up with: MOZILLA_HOME=/usr/local/netscape netscape (where instead of /usr/local/netscape, you would actually put in whatever directory you used - you seem to indicate that you put netscape in /usr/local/netscape which is why I used that dir.). If this works for you, you may consider replacing the symbolic link that's invoking netscape (you can find out where this link is with which netscape) with a little script that looks something like this: #!/bin/sh MOZILLA_HOME=/usr/local/netscape export MOZILLA_HOME exec /usr/local/netscape/netscape $@ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: what goes into .xinitrc and what into .xsession
Micha Feigin [EMAIL PROTECTED] writes: i tried to figure out while setting up the system whats supposed to go under the .xinitrc file and what into .xsession. Because I tried to use .xinitrc first to initialise stuff (like chosing window manager) but it won't work, so i put the things in .xsession and the most did, exept that every time that people tell me where to change configuration they tell me to put stuff in .xinitrc so i wondered. Maybe I'm doing something wrong. Thanx It is a bit confusing, isn't it. Here's the situation under Debian: If you start X with startx, then: first it tries to do the stuff in your .xinitrc; if you don't have one, it goes and invokes the system xinitrc, which on Debian goes and does what happens when you use xdm to start X. If you start X with xdm (the graphical login screen): It does the stuff in the system-wide Xsession. On Debian, this does nice things like suck in your ~/.Xresources file if you have one and then if you have your own ~/.xsession, goes and does that - if you don't have your own ~/.xsession, it tries to start an xterm and whatever you have set up as the default window manager for your system. The combined moral of this on Debian is: A ~/.xinitrc is pointless - if you put all your customizations into ~/.xsession and delete your ~/.xinitrc then you'll get the same environment no matter how you start X, and when using startx you'll get all the nice standard startup things from the system-wide Xsession. It's really the better way to go. Also how do i load the modules for fvwm2? they won't load for me with an error message saying that they should be called from inside fvwm (i tried t load them from inside fvwm2 but it didn't work) And how do i set a back groung pix under fvwm2 and afterstep (can it be jpeg?) The error message means to say that the modules should be called directly by fvwm2, (e.g. through a menu item or in a startup file). The easiest way is to install the menu package and then use the menus that appear when you click on the left button in the root window. (the root window is what is known to the rest of the world as the background - also note that you'll have to exit fvwm2 after installing menu) If you want to invoke a certain module with command line arguments, then you might try using the FvwmTalk module - this lets you send commands to fvwm2 as though it were reading them from a startup file. Also, if you've been getting non-Debian advice on configuring your system, someone may have told you to create a ~/.fvwm2rc file. Resist the temptation. Debian has a much better way to configure fvwm2 - read the /usr/doc/fvwm2/README.sysrc.gz for details. (Basically, create a directory ~/.fvwm2 and then put anything you want to add to the default setup into ~/.fvwm2/post.hook and all will be well) Finally, a background picture is easy with the Debian fvwm2 setup (I don't know about afterstep myself) - just put your background image into ~/.fvwm2/background.jpg. Or, to automatically size your picture to fit your screen, you can abuse the program that chooses background images and put the single line: -fullscreen /path/to/my/background.jpg into the file ~/.fvwm2/background.list - note that a ~/.fvwm2/background.jpg file takes precedence over a background.list file. You can add more lines to the background.list file to have it choose a random one each time you start X. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: how to make an intermediate machine receive mail while the destination is down?
Carlos Carvalho [EMAIL PROTECTED] writes: I need to shutdown our main mail server for hardware changes, and I'd like that another machine in our department receives the mail and hold it until the main server is up again, and then sends the stored mail to the server. How can this be done? The temporary holder can be our gateway/dns, for example, so that packets already travel through it. Carlos Well, you'll need to modify your MX dns record to point to the new destination - you'll then need to tell the gateway machine that's holding the mail explicitly how to route it, or else the machine holding the mail will try to send it to itself, and mail will get caught in a loop. How to do this depends on what mail software is on your gateway/dns machine. But why bother? If your main server is only going to be down for one or two days, then there shouldn't be any problem in simply not having something to hold the mail - if the sites sending the mail can't get through to your mail server, they'll just hold it until they can get through. Typically, what will happen is that people trying to send mail to your site will only even notice that something's odd if your server is down for three days (in that case, the person who sent the mail will get something from their mail server saying This message has been undeliverable for 3 days; we'll keep trying for another 9 days), and all the mail will get through if your main server is down for less than twelve. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: No E 14 yet
M.C. Vernon [EMAIL PROTECTED] writes: I too like fvwm2, but I can't configure the menu - I've RTFM and looked at the config pages, but I can't add things to sub-groups of the main menu - whatever I try just ends up at the bottom of the main menu (i.e. the `root' menu that comes up first when you click the mouse). I guess you must have got this sussed (I know a lot of packages auto-add themselves to the menus in the right place, but I'd like to add some more, and the big config file has do not edit all over it :( ) The problem with RTFM is that sometimes it's hard to figure out which manual it is one is supposed to read. Hooking into the auto-generated menus is best done through the menu package - point your favorite web browser to file://localhost/usr/doc/menu/html/index.html and read the section on How a user can override the menus. The advantage of this is that your new entries would appear in other menus should you choose to change window managers. (You can also make certain menu entries fvwm2 specific, but even then they'll appear in things like fvwm95 which can, in theory, do anything fvwm2 can) Alternatively, you could in your post.hook have code like: AddToMenu /Debian/XShells Super-user xterm Exec ssh -l root localhost xterm -title Root Shell -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: xemacs splits mail
Stefan Gödel [EMAIL PROTECTED] writes: I'm using VM. Does this really matter? I always felt that reading and writing / sending of mail are very distinct from each other (at least in xemacs). I'm using the usual mail mode with mime extensions to write my mail. My variable help gives me: `mime-editor/split-message' is a variable declared in Lisp. Value: nil Documentation: *Split large message if it is non-nil. [tm-edit.el] I have (setq mime-editor/split-message nil) in my ~/.emacs file. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: jdk1.1-runtime
Ed Cogburn [EMAIL PROTECTED] writes: What you are essentially trying to do here is to upgrade from libc5 to libc6 (the JDK was compiled for libc6). This is not a trivial task. It is the heart of the difference between Debian 1.3 and 2.0. There is a shell script called autoup.sh in the developer's corner on http://www.debian.org that can automate the worst of the upgrade (it will upgrade via ftp). After that, you can use dselect to finish the upgrade (also via ftp). Be warned though, it took me the better part of a day (~10 hrs) to download everything for this, although a minimal install should take less. The ftp location for hamm is 'ftp.debian.org/pub/debian/dists/hamm/main'. Although I was using the jdk1.1 for a long time before upgrading to libc6 - the old jdk1.1 package allowed one to install it on a bo system, which was most appreciated at the time. I still have the 1.1.3.v2-1 .deb packages if someone wants them (these can be installed on a libc5 system) - note however that there may be bugs the maintainer has fixed in recent packages. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Q: How to get class String from libg++272 with Debian2.0
[EMAIL PROTECTED] writes: On Fri, May 29, 1998 at 02:31:29PM +0200, Waldemar ¯urowski wrote: I switched from Debian 1.3 to Debian 2.0 while I was writing a program for my final exam of my study. I used in many places class String, and class which implements regular expression. When I did upgrade, I didn't check if everything had compiled ok. First of all, I had to use EGCS g++, which uses stdc++2.8, as g++ from wasn't anymore in GCC package. It seems there isn't any String class in stdc++-lib. I know that String _is_ in libg++-dev, but stdc++-dev and libg++-dev conflict which each other. I'd like to go back to g++ from GCC but I don't see it anywhere. SNIP I've built libg++2.8 and libg++2.8-dev packages for use with egcs's g++. Those packages are currently still in the Incoming directory, so you'll have to fetch them from a mirror of Incoming (see http://www.debian.org/devel/incoming_mirrors.html for a list) for now. Yay! I've been waiting for this; I know, I'll migrate my code eventually... The fastest way I found to getting to where I could compile old g++ source code was to install the altgcc and libg++-altdev packages; I know, it's not the best solution, but it was the one I could set up quickest at the time. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: WindowMaker | Menu? - Xterms
[EMAIL PROTECTED] writes: Anyone know how to get something like, xterm -bg black -cr green -fg white to be the command that gets spawned from WindowMaker's XShells-Xterm, instead of the ugly default xterm? Thanks in advance, Timothy Three solutions: 1) decide that the default xterm isn't so ugly after all. 2) Modify Xresources to make the xterm that appears with xterm be what you want; putting the following into ~/.Xresources (and either restarting X or running xrdb -merge ~/.Xresources) will do that: xterm*background: black xterm*foreground: white xterm*cursorColor: green (You can put these lines into /etc/X11/Xresources to affect every account on your machine) 3) Modify the menu entry - to do this, you should (as root) copy /usr/lib/menu/default/xbase to /etc/menu/xbase and then edit it, changing the xterm to xterm -bg black -cr green -fg white so that a portion of that file reads: ?package(xbase):needs=x11 section=XShells \ longtitle=Xterm: terminal emulator for Xtitle=Xterm\ command=xterm -bg black -cr green -fg white You should then (as root) run update-menus. You can also (if you want to make this change just for your account, and not for all the others on your system) put the modified xbase file into ~/.menu/xbase and run update-menus as your normal (non-root) user. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: jdk1.1-runtime
-BEGIN PGP SIGNED MESSAGE- I managed to find the old jdk 1.1.3v2 .debs which can be installed on bo (libc5 systems). I've put them into my public_html directory on master.debian.org - note that although I am a debian maintainer, I do not maintain nor never have maintened the jdk packages. md5sums: c875ec46c914747fd24d0e4e034c8101 jdk1.1-dev_1.1.3.v2-1.deb e0cb8617175008fc6a09ccfbc0c52840 jdk1.1-docdemo_1.1.3.v2-1.deb 630b752a48d7842b53dce47cb7a9920a jdk1.1-runtime_1.1.3.v2-1.deb Each of these are available via http: http://master.debian.org/~fizbin/jdk1.1-dev_1.1.3.v2-1.deb http://master.debian.org/~fizbin/jdk1.1-docdemo_1.1.3.v2-1.deb http://master.debian.org/~fizbin/jdk1.1-runtime_1.1.3.v2-1.deb -BEGIN PGP SIGNATURE- Version: 2.6.3a Charset: latin1 Comment: Processed by Mailcrypt 3.4, an Emacs/PGP interface iQCVAwUBNW8knxveYt4Z3sD9AQFllQQAjBCzKiye8QwE6UoRisnmZNV3zPeeMd4u mUflbfFU8koHhfyL5A5N1Edkukt3S8byFpGGLf6HvSkRYoWsSgTUSSsZy/wLtYJZ 26/QN3jJMQLm90fTClX82BBG0vGRUma8Uzr4mdqYQZHOCBwSj/f9GFWNKlMVSfXd eFc27gLnfqk= =ADGE -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: WindowMaker | Menu? - Xterms
[EMAIL PROTECTED] writes: Cool! Thanks alot! Can you tell me where is the documentation for this .Xresources file? man .xresources, man -k xresources etc don't have anything. :( Timothy Three solutions: 1) decide that the default xterm isn't so ugly after all. 2) Modify Xresources to make the xterm that appears with xterm be what you want; putting the following into ~/.Xresources (and either restarting X or running xrdb -merge ~/.Xresources) will do that: xterm*background: black xterm*foreground: white xterm*cursorColor: green (You can put these lines into /etc/X11/Xresources to affect every account on your machine) The real documentation? It's in the Xlib documentation published by the X consortium. Trust me, you don't want that. The basic format of the file is as in the lines above - my ~/.Xresources file looks like this: *scrollBar: true ! make everything have a scrollbar xterm.vt100*translations: #override \ KeyBackSpace: string(\177) \n\ KeyDelete: string(0x1b) string([3~) \n xterm*background: bisque3 xterm*foreground: black netscape*blinkingEnabled: False DctrlIcon*geometry: +1092+670 pmon*geometry: +1052+670 Blank lines are ignored - exclamation points begin comments. A particular application obeys resources if they begin with * or with the application's executable name or with the application's class name (an application's Class name can be found with 'xlsclients -l'). What each resource does is application specific. The xterm man page has several good examples of xresource lines. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: why can't i compile things ???
Mario Filipe [EMAIL PROTECTED] writes: Hi Yesterday i (think i) did a mess of my system. I accidently made a strip * on a lib directory. Today when i try to compile anything I always get the following : gcc -O -o poster poster.c -lm ld: warning: cannot find entry symbol _start; defaulting to 080485e0 Can anyone please expalin me how to fix this. Thanks You need to re-install at the very least the ldso package. (Packages usually need to be re-installed with dpkg, not dselect - just say dpkg -i /path/to/file/ldso-what.ever.deb) You should probably also re-install libc6 (assuming you have a hamm system; libc if you're still using bo, aka 1.3). -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: changing default screen manger in x
Rev. Joseph Carter [EMAIL PROTECTED] writes: On Mon, May 25, 1998 at 10:04:38PM +0200, Martin Bialasinski wrote: Make sure you have allow-user-xsession in /etc/X11/config, then create the file .xsession in your homedir. It should be something like: xterm # start a xterm xterm # start a second one fvwm2# start the windowmanager. Note the missing those should be exec'd, ie exec xterm , etc. WHAT? First off, not using the exec anywhere is _just_fine_. Some people don't like not using that because then there's an extra process floating around until you exit X. If, however, you do decide that you can't afford the (rather small) overhead, or you decide that it's messy to have extra processes just floating there, only the last line (the fvwm2 line) should have exec at the front. If you want to do any cleaning up after you exit X, you need that process hanging around and so can't use exec. exec command is a bit non-sensical. It does nothing different from command and just spreads confusion about what should be in an .xsession file. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: I can't get tar to exclude files!!
Admaster Communications [EMAIL PROTECTED] writes: I'm running Debian GNU/Linux 1.3 and I've tried many permutations of tar -zcvf test.tar * --exclude=leave_me_out.txt The file I ask to be excluded is reliably INcluded!! What is the correct syntax??? The problem is that --exclude applies only to files tar finds by descending directories, not to files mentioned explicitly on the command line. When you use *, that's the same as typing every file name explicitly (since the shell does not pass tar the *, but instead passes it each file's name - try 'echo *' some time to see what the shell does); therefore, --exclude doesn't apply. The following appears to do what you want - be aware, however, that this tar archive contains ., so when extracting it it will change the permissions of the current directory to the permissions of the directory when the tar archive was made: tar -czvf ../test.tar.gz --exclude='*exclude*pattern*' . (Note the ' marks and the final period - also note that test.tar.gz has to be in a different directory, or it will try to include itself - nasty recursion) Note that it's usually bad form to create a tar archive of files without a directory. That is, if the directory /home/martind/work/vander contains the files I want to make a tar archive out of, it's better to do: cd /home/martind/work/; tar czf vander.tar.gz vander than to do: cd /home/martind/work/vander; tar czf ../vander.tar.gz * The first method will mean that when one extracts the archive, all the extracted files go into a vander/ subdir. The second method means that all the newly extracted files go into the current directory, where there may already be much stuff. Note that if you use the first method, a --exclude will work as you might expect it should. If you really must have a tar archive with files in the current directory, and don't want to have . in the archive, then try: (this should all be one long line) find * -prune \! -name '*exclude*pattern1*' -print | xargs tar czvf testtar.tar.gz --exclude='*exclude*pattern2*' The difference between '*exclude*pattern1*' and '*exclude*pattern2*' is that the first pattern excludes files in the current directory and the second excludes files in subdirs. of the current directory. For example, if I had the following files on my system: /tmp/tartest/ /tmp/tartest/bar /tmp/tartest/tbark /tmp/tartest/frog /tmp/tartest/sounds/ /tmp/tartest/sounds/frogcall.au /tmp/tartest/sounds/dogbark.au /tmp/tartest/sounds/catmeow.au /tmp/tartest/froglegs/ /tmp/tartest/froglegs/recipe1.txt /tmp/tartest/froglegs/recipe2.txt and then I did: cd /tmp/tartest find * -prune \! -name '*frog*' -print | \ xargs tar czvf mytar.tar.gz --exclude='*bar*' I would then get (as the output of 'tar tzf mytar.tar.gz'): bar tbark sounds/ sounds/frogcall.au sounds/catmeow.au Note that the '*frog*' excluded the 'frog' file in the top directory and the 'froglegs' directory. Also note that the '*bar*' didn't exclude the bar and tbark in the top directory (the way the long command is written causes tar to be called with the arguments 'bar', 'tbark', and 'sounds' - these are the output of the find command), but did exclude the 'dogbark.au' in the sounds directory - this is because tar was handed 'sounds' as an argument, and descended into the directory itself. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Matrox Mystique 220 video card
Keith Alen Vance [EMAIL PROTECTED] writes: I recently installed a Matrox Mystique video card and was wondering what files I am missinf from my xfree86 and where I can get them. I don't have a browser setup on this computer so if you could give me ftp servers that would be great. Here is the error message I get when starting xwindows with afterstep 1.4 installed. I rerean the configuration utility and told it that I had a Matrox Mystique video. I did not select an accelerated server because I was told the matrox cards work with the svga server. For the Mystique 220 card, you really need a more recent Xfree86 than comes with Debian 1.3.1 (aka bo) - you have two choices: 1) upgrade to hamm (aka what will be Debian 2.0) 2) Install the XFree86 3.3.1 binaries yourself I think method 1) is preferrable, but you can find (by searching the mailing list archives for mystique from February) a description of how to do 2) as well. If you decide to do 1), be certain to look at http://www.gate.net/~storm/FAQ/libc5-libc6-Mini-HOWTO.html first. There's also a script that will do most of the upgrade for you somewhere - I think it's in the Developer's corner section of the debian web site. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Xwindows not working
Keith Alen Vance [EMAIL PROTECTED] writes: I have recently been screwing up my system and I am about ready to trash the whole thing and start over from scratch. I purchased a new video card Mystique 220 and I am trying to get it work with Xfree86 and AfterStep. I downloaded the files for Xfree86 3.3.1 and installed them and then setup afterstep 1.4. The problem is that if I do a startx if returns an error that it can't find afterstep.So if I go in to the .xinitrc file and comment the afterstep line out xwindows starts and then shuts down without any error messages. What I am saying is that it starts and then automatically shutdown right away. I really screwed things up when I was extracting the xfree86 files. I extracted them in the wrong directory at least once. So my hard drive is almost full. Is there a way I can delete Files by date. I would really like to totally back out of everything that I have done today and start over with a fresh kernel and the bare necessities. I am reluctant to just wipe my drive because I installed Linux via ftp and that takes awhile. Any suggestions would be great. Well, unfortunately extracting files usually gives the newly extracted files their old date, rather than the current one. However, what you might do is find out where the files went - you can do this with: find / -name 'XF86*' You can find out what a specific tar file extracts to with: tar tf whatever.tar or tar tzf whatever.tar.gz Then, go to the directory where you were when you did the incorrect extraction and try something like: tar tzf whatever.tar.gz | xargs rm -i (you may have to be in the directory above where you did the extraction, depending on how the tar file is built) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Why fetchmail didn't download my e-mail ?
Ionut Borcoman at musa [EMAIL PROTECTED] writes: SNIP bash-2.01$ fetchmail fetchmail: 1 message for borco at mail.mailbox.ro. reading message 1 of 1 (2579 bytes) fetchmail: SMTP error: 501 @mail.mailbox.ro : colon expected after route fetchmail: SMTP error: 501 ionut : sender address must contain a domain fetchmail: SMTP transaction error while fetching from mail.mailbox.ro bash-2.01$ In /var/log/exim 1998-05-20 17:29:35 unqualified sender rejected: ionut H=localhost (debian.borco.net) [127.0.0.1] (ionut) The problem isn't with fetchmail exactly, it's with exim - What appears to be happening is that error messages have no envelope sender (presumably to prevent stuff from being thrown back at them - there's no point in generating an error message for an error message), so fetchmail just uses your name as the envelope sender. Exim, however, doesn't want to accept your unqualified (unqualified here means a name such as ionut and not [EMAIL PROTECTED]) username as a sending address. The way to tell exim to behave more like less paranoid mailers and accept this unqualified address is to put the following in /etc/exim.conf: sender_unqualified_hosts = localhost Hmmm... I wonder if I've been missing some error messages because of this... (I didn't have this line in my exim.conf until now either). -- I need a sig... -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Overriding fvwm95 buttons in the post.hook?
Douglas Bates [EMAIL PROTECTED] writes: I have read the documentation on fvwm95 and on FvwmButtons but I am still at a loss to decide how I would override definitions of buttons in /etc/X11/fvwm95/system.fvwm95rc What I would like to do is to use rclock rather than xclock for one of the buttons. (rclock can be configured to alert you to incoming mail so a single window can do double duty of displaying the clock and an xbiff-like indicator of incoming mail.) If I add the lines *DebianFvwm95Buttons(Title rclock, Icon clock.xpm, \ Swallow(UseOld) rclock 'Exec /usr/bin/X11/rclock -bg \#c0c0c0 \ -geometry -1500-1500 ') to my ~/.fvwm95/post.hook file I get the rclock in addition to all the other buttons. What I would like to do is to remove the other buttons and then add the ones I want. Is this possible using the post.hook file? Well, I'm not entirely certain it is - this is one of the reasons I really need to get off my but and package up something like Redhat's configuration scheme. Actually... (checking the fvwm95 man page) Yes, it is possible. What you need to do is say: DestroyModuleConfig DebianFvwm95Buttons This wipes out all the default buttons - you can add the ones you want back in. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: pppd 2.3.5: peer refused to authenticate problem
Shaleh [EMAIL PROTECTED] writes: The answer is easy. The new ppp's set auth on by default. They did not use to do so. Edit the provider script and add noauth near the end. All will be happy. To be a bit more explicit - make certain that the keyword noauth is in /etc/ppp/peers/provider. Also, note that you no longer need to have your /etc/ppp/peers/provider (which is where the contents of bo's /etc/ppp.options_out got dumped) all one line. (In fact, I don't think you ever needed it all in one line, but now you certainly can use multiple lines); mine looks like: # The chatscript (be sure to edit that file too) connect /usr/sbin/chat -v -f /etc/chatscripts/provider bsdcomp 15 crtscts mru 2000 mtu 2000 defaultroute noipdefault /dev/modem 38400 modem persist noauth I find the file much more readable this way; not that it changes too often, but it just looked messy before. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
modprobe log entries
Does anyone know how to have the modprobe in hamm display more descriptive log messages when it can't load a module? I've been getting messages in my log like: May 14 13:27:41 cush modprobe: can't locate module and I've been trying to track them down, (they appear whenever netscape is started) but it would really help if I knew what module it is that modprobe can't locate (as a test, doing 'insmod IMnothere' gives the identical log entry - no clue of what module was asked for). I seem to remember that modutils from bo would actually tell you in the logs what was being looked for. (Can someone running bo confirm this for me? That is, that the unfound module is mentioned in the logs, not just as output of insmod.) I'm running hamm, modutils version 2.1.85-10. And just to head off some of the 'try this and see if it works' replies, yes I do have alias net-pf-3 off alias net-pf-4 off alias net-pf-5 off in my /etc/conf.modules. Hmmm... this may be a bug against modutils, as I'm almost certain that bo gave you a clue in the logs. -- Daniel Martin, who doesn't really have a good sig yet. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: modprobe log entries
Daniel Martin at cush [EMAIL PROTECTED] writes: Does anyone know how to have the modprobe in hamm display more descriptive log messages when it can't load a module? I've been getting messages in my log like: May 14 13:27:41 cush modprobe: can't locate module and I've been trying to track them down, (they appear whenever netscape is started) but it would really help if I knew what module it is that modprobe can't locate Well, I don't know what I thought I was doing before, because modprobe appears to still give full module information when it can't find a module; it's just that in this case the module being asked for is the empty string (); kerneld is calling modprobe with something equivalent to: /sbin/modprobe -k -s '' So maybe this is a bug against kerneld, as I can't imagine a normal situation that would cause it to ask for an empty module. And I have no idea what module binfmt-0 might be; it doesn't appear as one of the aliases on my machine's modprobe -c. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: xdm and shutdown
Jorge Daniel Ruckj [EMAIL PROTECTED] writes: Hi all. How I do a shutdown if I use xdm? Thanks Are you aware that even while using xdm one can press Ctrl-Alt-F1 to get to a text console? I find a surprising number of linux users unaware of this fact. From the text console, you can do a shutdown as normal. A while ago (within the last two weeks, I think) someone posted some nice tcl/tk code that one could add to some of xdm's startup scripts to get a little shutdown button to appear on the xdm startup screen. I'll see if I can go dig it up... -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: help w/ ttysnoop
Shaleh [EMAIL PROTECTED] writes: Can someone please help me setup ttysnoop?? I have set up the entry in inetd.conf. /etc/inetd.conf telnet stream tcp nowait root/usr/sbin/tcpd /usr/sbin/in.telnetd -L/usr/sbin/ttysnoops /etc/snooptab * socket login /bin/login When I telnet to my box, ttysnoops appears to not spawn. Any help appreciated. Have you told inetd to look at the new config. file? You do this with: (as root) killall -HUP inetd Or I suppose you could reboot, but that's win95-ish thinking. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: off-topic autofs permissions
[EMAIL PROTECTED] writes: Hi! I guess this it not a debian related problem, but: I am setting up a small Network with a linux server and some linux, win95 and nt-workstation clients. The clients shall be able to access the cd-rom and zip-drive on the server using autofs. The problem is that the zip drive is mounted as drwxr-xr-x root.root /misc/zip, so windoze clients can´t write on it. I RTFM, but did not find a way to set other permissons. Any ideas someone? There are two ways to do this - one is to make the zip drive have an ext2 filesystem instead of a dos one, but then you won't be able to use the zip drive on MS machines. The other is as follows: I assume that in your /etc/fstab you have something like: /dev/sda4/misc/zip vfat noexec0 0 (Or you may have type fat instead of vfat, or you may have some other options instead of noexec (like, say, defaults)) Anyway, what you need to do is make the following change to /etc/fstab: /dev/sda4/misc/zip vfat noexec,umask=0000 0 (If you just had defaults before, you can just replace it with the umask=000 bit) Then everyone will be able to write to /misc/zip. Also, anyone will be able to delete anyone else's files on /misc/zip, and you won't be able to tell who did it - this is because the fat/vfat filesystem can't keep track of who owns what file; attempting to make the directory sticky (like /tmp is) won't work. I suggest that it really might be best to have some other directory, which was set with permissions like /tmp that would be copied onto the zip disk every five minutes or so by a cron job. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: chmod not working
Alain Toussaint [EMAIL PROTECTED] writes: i was messing with some permission to directory but i think i found a bug in chmod,i try to run chmod 751 on my home directory (i put the content of the home directory on another partition because of lack of space on my root partition),but the permission is still drwxr-x--x,i even tried to run chmod u=rwx,g=rx,o=x on it to no avail,does all this qualify as a bug ??? I think that you must think that 'drwxr-x--x' means something different than it does. Seeing this permission in ls means that the user (owner) of the directory has read, write, and execute privileges, that people in the same group as the owner have read and execute privileges, and that anyone else has only execute privilege. Did you perhaps mistype the above and leave out the word not somewhere? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Unidentified subject!
(I've replied to debian-user instead of debian-devel because this really belongs on -user) Ian Keith Setford [EMAIL PROTECTED] writes: Yo- I would like suggestions and input on how to sell Debian in the sense of Debian versus RedHat, FreeBSD, or any other distribution. I will attempt to persuade a small committee of faculty to install Debian on a yet to be purchased machine. The machine is an experiment at my university to see if a select group of students can successfully administer a machine. I have my own views of why Debian is better and honestly I love it to death but the marketing and X-based configuration of RedHat seem to be weighing against me. I am also aware of a notion that FreeBSD is somewhat more secure than linux which I must also combat. Well, I can only provide some anecdotal evidence - my undergrad. school recently (in the past year) switched the student-admisitered unix box from an ugly slackware version to FreeBSD. While both, as far as I can tell, are administration nightmares (though the people managing it now seem to be making saner decisions than the old slackware group did, so it's not as bad), at least it was possible to get X working on the box when it was Slackware. This has not been possible so far for the FreeBSD folks. Also, last I looked at FreeBSD it was still using some old X - that is, earlier that X11R6.3. The consequence is that X windows on FreeBSD was _much_ less secure as anyone who could telnet in could grab a picture of the display (or worse). That, however, may have changed. So if X is a priority, don't go with FreeBSD. Where FreeBSD's more secure reputation comes up is that occasionally bugs are found in the Linux networking kernel code that make it possible for people to write programs which take the machine down. These are fixed and patches are usually distributed within a week - often before many people even know the bug exists. FreeBSD rarely has this problem. This is in my experience the only time the reputation of FreeBSD being more secure is justified; also, it's a relatively minor worry, what with recompiling the linux kernel being as easy as it is. Finally, I'd avoid FreeBSD for the same reason I'd avoid slackware - no packaging system. On a student-run box (and therefore, presumably administered by volunteers and/or people with less than 5 years experience as unix sysadmins), this is a definite priority. RedHat's only advantage over Debian that I can see is in the ease of initial setup - and there, only in the ease of initial X setup. (which is, IMHO, a major flaw in Debian 1.3.1 - X Setup is poorly documented, or at least it was when I did it - maybe the FAQ-O-Matic has taken care of things). RedHat is a pain to upgrade and a royal pain to administer remotely. With Debian, it's a snap. (The bo-hamm upgrade notwithstanding, which is easy if you actually read the documentation beforehand). Please suggest the best hardware configuration for a machine that will host thousands of users and be up 24/7. I suspect that an i386 machine would be most suitable only because I am unaware of the stability or availability of Alpha/Sparc ports and due to the lesser price of i386 hardware. Sure - Pentium or whatever. The most important thing is to have gobs of memory - 64 meg minimum. Also, if you buy the hardware in components yourself, make certain you have enough fans. (Some Cyrix chips are known to overheat - if you have the money, there's nothing wrong with buying genuine intel). Please point me towards any information that might help me get Debian on such an important machine at my university. The machine that am I asking for advice in building is a replacement machine for the machine in which this e-mail is being created. I am sorry if this message seems somewhat convoluded but I feel very strongly about having Debian an the operating system on this computer. I lack the advanced programming skills(I receive my degree in Finance in a few months) to be a Debian developer but I sincerely desire to help promote and help the Debian project. On a side note, with whom do I need to correspond with to discuss the use of the Debian logo and name? If anyone was wondering, I attend the University of North Texas located in Denton, Texas (30miles north of Dallas/Ft.Worth). It is the 3rd largest university in Texas. Please feel free to reply directly to me or to the list! Thanks again, -Ian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
modprobe log entries
Does anyone know how to have modprobe display more descriptive log messages when it can't load a module? I've been getting messages in my log like: Apr 22 16:38:52 cush modprobe: can't locate module and I've been trying to track them down, (they appear whenever netscape is started) but it would really help if I knew what module it is that modprobe can't locate (as a test, doing 'insmod IMnothere' gives the identical log entry - no clue of what module was asked for). I seem to remember that modutils from bo would actually tell you in the logs what was being looked for. (Can someone running bo confirm this for me? That is, that the unfound module is mentioned in the logs, not just as output of insmod.) I'm running hamm, modutils version 2.1.85-10. And just to head off some of the 'try this and see if it works' replies, yes I do have alias net-pf-3 off alias net-pf-4 off alias net-pf-5 off in my /etc/conf.modules. Hmmm... this may be a bug against modutils, as I'm almost certain that bo gave you a clue in the logs. -- Daniel Martin, who doesn't really have a good sig yet. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]